Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
G-o-ve_rn_m-en-tA-cc'-es-sio_n_N_o.-----.--, crNITR
"illll
CH LIBRARy~on
Page
5. Report Dote
May 1993
DETAILING FOR STRUCTURAL CONCRETE
7. Author(s)
~A_u-:-s-ti_n_,_Te_x.~a_s_-:7-:8_7_0_5_--:-2-:-6-:-5-:0---------------l
12. Sponsoring Agency Nome ond Address
Final
Study conducted in cooperation with the U.S. Department of Transportation, Federal Highway
Administration. Research Study Title: "Reinforcement Detail Design in Structural Concrete"
16. Abstract
This report is the final report in a series which investigates the applications of strut-and-tie modelling
for typical details in structural concrete bridges. It summarizes the state of the art of strut-and-tie
modelling and presents specific recommendations for choosing the critical dimensions and carrying out
detailed computations using such strut-and-tie models. Separate sections treat the overall modelling
and detailing process, checking compression struts, detailing tension ties, evaluating TTT, CCC, CCT
and CTT nodes, and incorporating prestressing forces. The report includes a series of examples
showing application of strut-and-tie models in detailing deep beams, corbels, anchorage zones, dapped
ends, openings, and pretensioned beams. In addition, a number of detailing aids are included in an
appendix.
Unclassified
Form DOT F 1700.7 {8-72)
Unclassified
Reproduction of completed page authorized
316
22. Price
IMPLEMENTATION STATEMENT
_.)
iii
The contents of this report reflect the views of the authors, who are responsible
for the facts and the accuracy of the data presented herein. The contents do not
necessarily reflect the official views or policies of the Federal Highway Administration.
This report does not constitute a standard, specification, or regulation.
There was no invention or discovery conceived or first actually reduced to
practice in the course of or under this contract, including any art, method, process,
machine, manufacturer, design or composition of matter, or any new and useful
improvement thereof, or any variety of plant which is or may be patentable under the
patent laws of the United States of America or any foreign country.
Research Supervisors
iv
PREFACE
This is the final report in a series of three reports which investigated
applications of strut-and-tie modelling for typical details of structural concrete bridges.
Research Report 1127-1 looked more specifically at the problems of shear and
diagonal tension in the negative moment regions of precast girders for use with dropin spans. Research Report 1127-2 summarized a series of tests of typical details used
with dapped beams and several different types of nodes. This report (1127-3F)
presents a summary of the basis for strut-and-tie model use in detailing structural
concrete and includes a series of illustrative examples.
~-
_)
l_
L_
~J
'
TABLE OF CONTENTS
Chapter 1 Introduction
1.1
1.2
1.3
1.4
General Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Summary of the History of Detailing
.................. .. .. ... ..
Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Scope
.. .... ........ ........ .. ................. ............
5
24
27
_)
Chapter 2 Background
'_j
2.1
2.2
2.3
2.4
2.5
' 1
2.6
2.7
2.8
; J
2.9
............................................... .
29
29
33
38
41
45
63
69
71
84
88
105
107
112
119
124
128
133
135
135
136
136
139
3.1
3.2
3.3
vii
3.4
4.1
4.2
4.3
149
150
151
157
160
162
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
Detailing Aids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Typical Examples of Detailing Aids . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.3.1 Load Near Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.3.2 Corbel Projecting from a Column . . . . . . . . . . . . . . . . . . . . . . .
4.3.3 Deep Beam with a Hole . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.3.4 Dapped End Beam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.3.5 Anchorage Zone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.3.6 Pretensioned Beam with Eccentricity . . . . . . . . . . . . . . . . . . . .
165
168
169
171
189
205
226
243
257
271
References
273
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285
viii
LIST OF FIGURES
Figure
1.1
1.2
1.3
1.4
1.5
1.6
1.7
Page
Typical examples of strut-and-tie models
6
8
13
15
19
............................... .
20
20
21
23
34
1.8
1.9
1.10
1.11
2.1
2.2
2.3
2.4
2.5
2.6
2.7(a)
2.7(b)
2.8
2.9
. J
2.10
2.11
2.12
2.13
2.14
...........
........................
.............................. .
................................. .
35
37
40
42
.................................. .
43
................................... .
44
.......................................... .
44
47
............................................. .
51
51
53
53
55
Distortional effect
ix
_j
.............. .
49
2.15
2.16
2.17
Mohr's circle
..................................................
Proposed strut-and-tie model for shear behavior
..................
Comparison of test results with the theoretical approach of
predicting the diagonal compression strength
.....................
68
2.18
70
2.19
71
2.20
74
......................................... .
76
58
66
2.21 (b) Variable dimensions for geometry of the bearing and loaded plate
77
2.22
........................
78
2.23
83
2.24
86
2.25
2.26
86
87
2.27
Pretensioned beam
91
2.28
.............................. .
93
2.29
96
2.30
99
99
2.32
2.33
2.34
99
(c) Eccentric pretension force .................................. . 100
(d) Fully plastic strut-and-tie model
............................. . 100
(e) Strut-and-tie model for the end "D" region .................... . 100
Strut-and-tie model for prestressed concrete beam
with curved tendon
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
Approximation for radial compression component of curved tendon
103
Prestressed beam with parabolic tendon
. . . . . . . . . . . . . . . . . . . . . . . . . 104
Width of the compression-and-tension chord
. . . . . . . . . . . . . . . . . . . . . . 104
2.35
Types of nodes
2.36
CCC nodes
2.31
2.37
l_
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
108
108
109
'
,_
I_
2.38
CCC node
2.39
2.40
2.41
2.42
2.43
2.44
2.45
2.46
110
............. . 111
. . . . . . . . . . . . . . . . . . . . . . 115
116
............................. . 117
Dependency of the efficiency factors for CCT node
............... . 118
2.47
2.48
121
2.49
124
2.50
125
2.51
126
2.52
127
2.53
2.54
.......................... . 130
Lateral pressure and the distance "e" to the reinforcing bar
........ . 131
2.55
2.56
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
138
142
143
146
146
147
148
149
152
3.9
3.10
3.11
xi
, __j
. . . . . . . . . . 114
3.12
3.13
156
158
3.14
3.15
159
160
3.16
161
3.17
4.1
Types of nodes
4.2
4.3
4.4
4.5
.......................................... . 183
4.6
187
4.7
Reinforcement layout
4.8
4.9
4.1 0
4.11
4.12
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
193
4.13
194
4.14
4.15
195
196
199
203
204
206
4.16
4.17
4.18
4.19
4.20
4.21
.............................. . 163
............................................... . 169
170
................ . 173
...................... . 174
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
191
. . . . . . . . . . . . . . . 192
207
211
211
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
. . . . . . .. . .. . . . . . . . . .. . . . . .. . . . . . . . . . . . . . . . . 214
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 6
xii
,-~,
4.21
4.22
4.23
4.24
4.25
4.26
4.27
228
230
4.28
230
4.29
231
4.30
232
4.31
4.33
4.34
245
4.35
246
4.36
247
4.37
4.32
4.38
4.39
4.40
4.41
4.42
4.43
227
227
228
253
Strut-and-tie model for example: anchorage zone
256
Reinforcement layout for anchorage zone
................................. . 260
Pretensioned beam: geometry
Strut-and-tie model for prestressed concrete ...................... . 261
(a) Transfer length forces
..................................... . 261
(b) Model with tendon eccentricity effects
. . . . . . . . . . . . . . . . . . . . . . . . 261
(c) "D" region at end
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262
Reinforcement layout for pretensioned beam
269
xiii
LIST OF TABLES
Table
Page
1.1
2.1
2.2
2.3
2.4
2.5
2.6
2.7
2.8
2.9
2.10
2.11
2.12
2.13
2.14
2.15
2.16
2.17
2.18
2.19
4.1
4.2
................. .
Statistical data for confined concrete strength with various test data . . .
Statistical data from Figure 2.23 for confined concrete
with an efficiency factor Ve = 0.5 + 15/(f c)0.5
.......................
Statistical data for confined concrete with
an efficiency factor Ve = 0.5 + 20/(f c)0.5
Friction and wobble coefficient
Loss of prestress
Design steps for 'B' regions of prestressed beams
using the strut-and-tie method
..................................
Statistical data from Figure 2.45 (CCT node)
......................
General information about the tested CIT nodes
Statistical data from comparison in Figure 2.55
Statistical analysis from Figure 4.15 omitting unreinforced specimen . .
Statistical data from Figure 4.31
.................................
16
32
57
60
62
64
67
68
69
70
81
82
82
93
94
102
118
122
131
196
233
~~--'
xiv
~J
SUMMARY
XV
\_
CHAPTER 1. INTRODUCTION
1.1
General Introduction
A structure must be safe, serviceable and durable during its lifetime. In
limit state and can be consistently treated by well recognized similar analysis
techniques at the serviceability limit state. It will greatly reduce confusion and
possible
error
if
consistent
treatment
_,.
..,--.-.--.
)
c.
Corbel
d. DeepBeam
Figure 1.1:
,.. ",.,....... ..
.. .,._
'
~
+++ ~
' .,. .,1~+++.-t-'"
Jr
#I
4 . . . . . _,..,. ....
+ ......-+'- '*'
,,.1.-t-+..C.+-t-:\..P'
.6r r r .,.. ~
.,.T.,. .,. + + + .,.. "" "
"
....
.,.
!.
t
..1++++++1.
' .9
c ~ c 9 .
...........
,,
:
. ........ .
t 1 t
tension
lltl
,,
11 1
lllltl
1111
til
w(x) = P I (b h)
L__,
Figure 1.2:
1.2
Corner
L_,
L ,
l~ i
Figure 1.3: Details of special concern that may exist in actual structures
_j
' .. J
7
Since all parts of a structure including the discontinuity regions are of similar
importance, an acceptable design concept must be based on a physical model
with a logical understanding. Truss models, because of their transparency and
adaptibility to many design situations, are seen as attractive alternatives to
empirical approaches for detailing structural concrete. Truss models for shear
design of reinforced concrete beams were introduced by Wilhelm Ritter [5] in
1899. Ritter introduced his model to dispel the idea that the main function of the
)
a.
--= ..
.. ....
- ID~
'r'
.. - c;J.~
--,.
. . .. .. .:.'*-
1111111111 11 I 111111111111
~~
P/2
Figure 1.4
(c)
P/2~
'_j
_j
10
In 1927, Richart [11] expressed the shear capacity of concrete beams with
vertical stirrups by an equation of the form:
where
v
r
av
s
b
fv
=
=
=
=
=
=
a/ (s b)
cross sectional area of web reinforcement
spacing of web reinforcing bars, measured at right angles to
their direction
width of beam
tensile unit stress in web reinforcement
factor which varies between 90 and 200 psi (depends upon the
the percentage of web reinforcement used and also on the
quality of the concrete).
This expression indicates that the computed stresses from the truss model were
lower then the measured stresses. The factor "C" was included to express
the additional mechanism for shear behavior, like aggregate interlock (friction),
dowel action etc. These basic ideas found wide use in American design
standards throughout the Twentieth Century. Since the majority of members
designed were subjected to only low or moderate shear levels, an empirical "C"
or concrete contribution (Vc) was introduced to supplement the truss model
capacity (Vs). The present US expressions [3, 4] for shear capacity are of the
pattern:
11
Continued use of the supplementary "V0 " term in US practice was encouraged
by more contemporary leaders such as Hognestad [12] who stated in 1951 that
if designs are made on an ultimate limit state basis the truss model will not result
in safe and economic structures under all conditions met in practice.
'- 1
( __ _j
_j
=
=
f I (A a)
generalized strain corresponding to ai
positive constant
1, 2 .. n
12
Starting from the yield condition and flow rules it is possible to derive the
theorems of limit analysis. The lower bound theorem states (see Fig. 1.5 (a)):
Solutions for the upper bounds are derived by equating the external work
done to the internal energy dissipation for the assumed mechanism. Upper
bound solutions are generally unconservative. The theory of plasticity states that
there is a unique and exact solution such that both the upper and lower bound
theorems are satisfied. The quality of a plastic analysis is dependent on the
constitutive equations used. These constitutive models of material behavior
define the yield condition which determines failure of the plastic model. The way
constitutive equations (most are empirically derived) are handled by the models
will be discussed in Chapter 2.
Various researchers, including Leonhardt [15], ROsch [16], Lampert [17],
Grob [18], LOchinger (space truss) [19], Muller (optimum inclination of the
diagonal members) (20], Neilsen et al. [21 ], Mitchell and Collins [22], Ramirez
and Breen [23] have worked to refine and expand the method so it is applicable
13
M1 =- F 1/3
M2 = + F 1/9
M1 = + 2 F l/9
M,
11s
1/3
1/3
__
""' :3Jl_
~
Mplastic
!13
1/3
Figure 1.5:
1/3
14
to shear, torsion, and the interaction of these actions, as well as bending. The
space strut - and- tie-model with variable angle of inclination of the compression
diagonals departs from the traditional truss model with 45 degree angle
diagonals. The angle is chosen such that in the field where failure occurs, both
the longitudinal and transversal (stirrups) reinforcement will reach their yield
strength. In this case a sufficient shear transfer by aggregate interlock across the
initial inclined cracks is assumed so that the concrete diagonals can reach their
final inclination under ultimate load. Due to the fact that such shear transfer
across a crack decreases with increasing crack widths (rough crack model [24,
25]), additional considerations become necessary. Hence, limits on the inclination of the concrete diagonals must be introduced. The model is valid in the
complete range of interaction between general bending, normal force, shear
force, and torsion. However, limits must be set in some fashion to preclude initial
compression failures. Recently, MacGregor and Gergely [26], Marti [27],
Schlaich et al. [28], Schlaich and Schafer [2] have published refined methods for
detailing structures using strut- and- tie- models. In the Canadian GSA-Standard
[29] the compression field theory, an idea somewhat similar to the strut- and- tiemodel, was introduced in 1984. Cook and Mitchell [30] published studies on
regions near discontiunities. The strut- and- tie- models were compared with a
nonlinear finite element study and test results.
The "Design and Construction Specifications for Segmental Concrete
Bridges" [31] introduces also the strut-and-tie-model as a design tool for areas
where the strain distribution is non-linear.
For prestressed beams with unbonded tendons, Kordina et al. [32]
compared theirtest results with the truss- model (Fig. 1.6(a)) and also with a tiedarch model (Fig. 1.6(b)). According to the truss model, the shear-carrying
capacity increases approximately linearly with the amount of web reinforcement,
whereas with the tied-arch model the shear-carrying capacity depends only on
the load-carrying capacity of the arch orthe tension chord. The comparison ofthe
15
test results with the truss model (45-deg truss was assumed) showed that the
shear capacity could be predicted best for unbonded prestressed beams.
Conversely, for a tied-arch model, an insufficient agreement with the test results
was obtained; only the compression-arch failure was considered as a failure
mode. Measurementsindicatedthatthestirrupstressesin beams with unbonded
prestressing do not differ in principle from comparable beams with bonded
prestress reinforcement, as far as the shear -carrying system is concerned.
I~
I
.
t
[fA~r,,f;!It~
l
fA
Ttnlbn bQ ~ lht
ti!Mir:lft c:hord
~~th ~~--------------~~~~
Forces cx:ti'lg at a truss Jc1W
Figure 1.6
16
Similar to the case of bonded prestress , the stirrup forces are reduced by a
part of the shear force carried by the concret~, including parts of the shear
force carried by aggregate interlock, the compression zone, and the dowel
action of the longitudinal reinforcement.
Recent studies by Hartmann, Breen and Kreger [33] compared the
truss model with test results for prestressed concrete girders using concrete
strength in the range of 12,000 psi. The 45 degree truss model gave very
conservative results when using a concrete efficiency factor of 0.5. Table 1.1
contains the statistical comparison of the analysis with different models.
Table 1.1:
Method
Experiment/theory: mean
standard deviation
ACI [2]
1.18
0.18
CSA [28]
1.74
0.78
Truss:
v e = 30 I (f'c) 05
1.72
0.26
Truss:
v e = 0.5 (f'c)
1.38
0.32
l _ _,
L___ ',
It is interesting that the traditional ACI- AASHTO shear expressions [3,4] provide
the closest agreement, once again indicating that the use of a supplementary "Vc"
term is important for economy in
Presence of precracking of the webs due to other loading patterns would greatly
affect and could substantially diminish any Vc contribution. One of the problems
in evaluating test data and comparing it with the truss model is to define the strut
width, strut angle and the efficiency factor of the concrete.
L.,-
-1+
17
Powers [34] studied prestressed girders with high strength concrete.
Evaluation of the test results and the comparison with the strut-and-tie-model
show that even for prestressed girders with high strength concrete, failure can be
predicted on the basis of a strut-and-tie-model. For the web-crushing failure the
limitation of the efficiency factor of the concrete is highly important.
Schaefer [35] and Castrodale [36] have shown that there is good agreement between the truss model and observed test results in both reinforced and
prestressed concrete beams subjected to different loading combinations of
bending and shear which extensively precrack the girders. Experimental results
of the behavior of reinforced concrete beams with various arrangements of
stirrups were compared with the strut- and-tie model by Kotsovos [37]. He
concluded that in general the strut- and- tie- model does not provide any detailed
information with regard to the strength and deformation of concrete.
Related work on this project by Barton [7] studied the application of a strutand- tie- model to beams with dapped ends. The various singular nodes that
may occur in the strut- and- tie- models were studied by Anderson [38] and
Bouadi [39]. Their detailed observations are summarized in the accompanying
report 1127-1 .
Noguchi and Watanabe [40] applied the strut- and-tie- model based on a
finite element study for the shear resistance mechanisms to beam-column joints
under reversed cyclic loading. The strut-and- tie- model gave good agreement
with the shear stress distribution model for all test specimens. Breen and Stone
[41], Burdet [42] and Sanders [43] investigated strut-and-tie-model approaches
based on elastic finite element studies and experimental tests, for the design of
post-tensioned girder anchorage zones.
18
The approaches of the various authors cited differ in the treatment of the
prediction of ultimate load and the satisfaction of serviceability requirements.
Schlaich et al. [28] proposed in general to treat the ultimate limit state and
serviceability in the cracked state by using the same model. This was to be done
by orienting the geometry of the strut- and- tie- model based on elastic stress
fields and by analyzing the resulting strut-and~tie-model structure following the
theory of plasticity. A computer based design approach based on these ideas
was developed by M. Schlaich [44].
The concept of a strut- and- tie- model can be used not only for statical
or geometrical discontinuities but also for other load transfer mechanisms like
anchorage provisions, dowel action and force transfer between concrete and
steel. Yankelevsky [45] described a truss model fort he force transfer between the
concrete and the steel by using static equilibrium and compatibility to relate the
forces. By knowing the axial force in the steel (a differential equation was solved
for the axial force in the steel), the bond shear stress was predicted and was of
an exponentially decaying form, maximum at the bar's ends and minimum at the
specimen's midspan (see Fig. 1.7). Another application of strut- and- tie- models
to details is the three dimensional truss model suggested forthe fracture behavior
,' __
of concrete. Rode [46] used a three dimensional truss model cube (Fig. 1.8) for
a computer simulation to study crack opening and crack growth. The model
conception is based on a 1941 idea of Hrennikoff [47] for the solution of linear
elastic continuum problems by a three dimensional framework method. This
model allows simulations on micro- and macro levels without altering the number
of elements. The basic cell is a truss cube with edge struts, surface diagonal
struts and space diagonal struts. The struts themselves behave linearly elastic
up to given strain rates. On exceeding the maximum tensile or compressive
r '
strain, the affected struts are removed from the system, representing cracks. A
l_'
19
I
I
I -
Compression In concrete
Compression in
concrete
Tension In bar
Compression In concrete
"""'' _.....
Figure 1. 7:
__ )
20
-edge strut
--- surface diagonal strut
--- space diagonal strut
Figure 1.8:
1.00
Vou/v:U
0.75
0.50
0.5
0.25
d= 24 n1m
fsy 440MPa
vtu [17]c 0.7v:.. [18}
* =Tsy
Tu
0.0
Figure 1.9:
0.25
0.50
0.75 Tu/Tu*
21
single strut represents the stress flow mainly through an aggregate particle,
another one represents the flow mainly through the mortar matrix, and a third one
is affected by the bond between matrix and aggregate. The strut parameters are
stochastically endowed with values by a computerized random number process.
All quantities of the strut parameters are normally distributed with a variance of
50%.
Soroushian et al. [48] and Vintzeleou [49] studied the dowel action with
regard to bond, tensile strength of concrete, and bar- to- stirrup interaction with
stirrup tension stiffening (see Fig. 1.9).
Different studies were done to investigate the shear transfer mechanism
based on aggregate interlock (Bazant and Gamabarova [50], Divakar, Fafitis and
Shah [51]). The two-phase model (Walraven [52]) has a rational formulation
based on a few assumptions..(See Fig. 1.1 0.)
=
cr_ =
(Jmc =
(Jwv
Jl
=:::
(Jmc (Aw + Jl Av )
(Jmc (~- Jl Aw )
0.394 f'c o.ss (matrix compressive strength)
0.4 (friction coefficient)
22
vmax =
vmax =
w
da
=
=
Gamabarova [53] compared the truss model with test data and found that
for thin-webbed I beams the truss-model is a quite conservative approach.
Therefore, more
account for shear design. This conclusion reemphasizes the findings of Talbot
[9] and Richart et al. [54] in the 20's and Hognestad [12] in the 50's that a
concrete term was needed to amplify the truss model for economical design of
lightly loaded members. Also Brandtzaeg [54] concluded in his theoretical
"analysis of stresses in a material composed of non-isotropic elements" that the
limiting value of shearing stress is the sum of the shearing strength of the
material and the coefficient of internal friction times the normal stress.
L __
~ -"~"-------~-~-----=='""-=..;;;,;,o,.........~~-.............lliiiioiilliiiioiiiiiioiiiiiiillillliliiillillliliiillillliliiillillliliiillillliliiillillliliiiliililiiiliiiiii-------Eil-"~'
23
There are several additional load-carrying mechanisms that can supplement the
basic truss model:
dowel action of the longitudinal bars (vertical)
aggregate interlock transfer forces across a crack (vertical and normal)
component of inclined prestressing tendons (vertical)
These campo nents of force transfer must either be neglected or treated indirectly
in the strut- and- tie- model. In lightly loaded members with low levels of shear,
such components are significant and some supplementary design mechanism
continues to be necessary for economy [55]. Currentdesigncodesand standards
work satisfactorily for typical "traditional" members such as uniform depth beams
with well distributed loading. Empirical solutions at supports have been developed to give good designer guidance for such cases. However when irregular
members are used, such guidance is limited. For example, the increased usage
of concrete bridge substructures and superstructures in highly congested urban
areas has caused increasing complexity in bent cap geometry, the introduction
of new cross section shapes and the increasing usage of precast, prestressed
beams cantilevered over a support to a hinge away from the support. Simply
supported drop-in beams are placed from one hinge to another hinge at the other
end of the drop-in beam. (See Fig. 1.11.)
24
This generally necessitates large notches or daps" being used at the cantilever
span end and the ends of the drop-in span. The many unusual bent configurations
in congested urban areas result in many highly loaded short bracket and corbel
applications in reinforced and post-tensioned concrete members. The increasing
usage of pretensioned concrete in unusual long span situations as well as in
massive bent caps creates a host of new applications of reinforced, pretensioned
and post-tensioned members and assemblages. Traditional code rules and
simple reinforcement patterns based on the simple span test specimens utilized
for experimental determination of so many of the ACI [3] and AASHTO [4] design
provisions do not provide guidance and are not applicable to many of these new
applications.
In order to avoid potentially serious strength or serviceability problems,
better guidelines for proportioning and locating reinforcement are needed. Such
guidelines should consider the full range of reinforcement from the passive action
of non prestressed bars to the active action of prestressing tendons, as well as the
case of mixed reinforcement (active and passive) which is becoming widely found
in post-tensioned concrete. Comprehensive detailing methodology and guides
are needed in practical detailing.
1.3
Objectives
-,
25
zones or regions where traditional approaches incorporating a concrete
contribution are inappropriate or uncertain will eliminate the economic penalty
occuring in general usage of strut-and-tie-models with lightly loaded members.
In the strut-and- tie model, the actual stress distribution within a structure is
idealized as a static force system consisting of the following basic elements:
struts:
columns, and slabs with tJniform loading patterns are only briefly addressed. This
study has as a basic objective the development of a consistent methodology and
an accompanying comprehensive detailing guide for structural concrete based
on use of refined strut- and- tie- models. It is hoped that the methodology and the
illustrative guide will help designers develop a clearer understanding of the
functioning of reinforcement and anchorage details in a wide variety of details in
concrete structures. It is envisioned that the designer will approach the detailing
of a concrete structure using strut-and- tie- models which may be based on an
equilibrium analysis of load paths, on detailed results from a linear finite element
analysis, or by analogy with a steel design procedure. After isolating
the
26
a suitable strut- and- tie- model to carry the applied loads and meet the given
boundary conditions. After selecting and analyzing a strut- and- tie- model, a
major concern are the nodal zones where inclined compression struts, vertical
stirrups, and longitudinal reinforcement intersect. The actual patterns of the
nodes and the limiting stresses in the nodes must be quantified before practical
implementation. Similarly, the allowable or effective compressive stresses in inclined compressive struts must be carefully evaluated.
The specific objectives of this overall study are:
(1)
(2)
(3)
I_
(5)
Objectives (2) and (3) have been reported in detailed in Reports 1127-1 and
,-
1127-2. This report summarizes the efforts to meet objectives (1) and (4) and
presents the detailing guide of objective (5). The guide should lead to more
consistent, constructible, economical and reliable details.
'
i '
27
1.4
Scope
. J
modeling
techniques as well as dimensioning of the struts, ties and nodes to illustrate the
method and facilitate its use. Recent studies about high strength concrete were
also included and design provisions evaluated. Suggested Design Specification
28
,--,
CHAPTER2
BACKGROUND
2.1
Concept Background
The strut- and- tie model is a limit analysis approach to the design of structural
concrete. More specifically, the strut- and- tie model is a static or lower bound
plasticity solution. Marti [27] explains that strut- and tie models represent a possible
equilibrium system of forces within a structure at its ultimate load. While the plasticity
theory behind the strut- and- tie model is quite complex [56], it is primarily used to
establish a rational basis for the method. For most practical applications, it is only
necessary to understand that a properly chosen and dimensioned strut - and- tie
model represents a lower bound (or conservative) estimate of the true capacity of a
structural element assuming other brittle failures such as stability or local crushing
are precluded.
Although development of detailed mathematical verification for the strut- andtie- method is unecessary to understand its application, awareness of the
assumptions is important. The most important of these assumptions are:
1.
2.
Crushing of the concrete struts should not occur prior to yielding of the
ties. This is prevented by limiting the stress levels in the concrete.
3.
4.
All external loads including post- and pretensioning forces are applied
at the nodes of the strut- and- tie model. In the case of distributed loads
and pretensioned strand loads, the model must be adequately
formulated to realistically represent the load distribution
5.
29
30
(2)
L..
31
3.
Slabs and shells are structures with two dimensions considerably larger
than the third with loads acting transverse to their plane or curved middle
plane. If they are predominantly uniformly loaded, they will essentially
consist of 8-regions. Strips taken along the principal moment directions
behave and can be treated therefore as linear structures.
If a structure contains 8-regions in substantial part, it is usually more
convenient to first determine its sectional load effects {M8 , MT' V, N) by use of
conventional elastic analysis.
For uncracked B-regions the internal stresses then can be determined
from the sectional load effects by use of cross-sectional values A, 18 , IT' and the
usual laws of mechanics (bending theory). In the case of high compressive
stresses the linear analysis of internal stresses may have to be modified by
replacing Hooke's law with one of the nonlinear material laws. For cracked Bregions the internal forces are generally determined from the cross-sectional
load effects by application of the standard truss or ordinary cracked reinforced
concrete theory. In 8-regions the use of truss models or of strut- and- tie- models
will often be more complex than required.
The forcepaths or the struts and ties of the 0-regions can be determined
from the loads applied to the D-regions by equilibrium analysis. If a structure or
member consists of only one 0-region, the analysis of sectional effects by a
conventional structural analysis may be omitted and the internal forces or
stresses may be directly determined from the applied loads. If the structure is
externally statically indeterminate, the internal compatibility of stresses should
be considered by first orienting the geometry of the model to the pattern of forces
indicated by the results from a conventional elastic analysis and then possibly
reorienting it thereafter according to tile major design intent: emphasis on
ultimate load capacity or on serviceability under working loads.
32
Table 2.1 :
Analysis usage
D-reg ions
B-regions
analysis
sectional analysis
direct
material behavior
linear or nonlinear
nonlinear
state I
= uncracked
sectional values
state II
=cracked
conventional reinforced
concrete analysis
or truss model
It is usually most convenient to orient the geometry of the strut- and- tie
models to the general pattern of load paths traced by the forces passing through
the member. These load paths can be determined from intuition, experience, or
in unusual cases by examining the elastic stress fields indicated by a finite
element analysis. Design of B-regions is accomplished by ordinary cracked
reinforced concrete theory or by using a special type of strut- and- tie model which
is generally termed the truss analogy. In the truss model for a simply supported
beam the upper horizontal chord represents the concrete compresssion zone.
The lower horizontal chord represents the main tension reinforcement The
stirrups of the beam are lumped together as the truss vertical members. Inclined
compression struts are used to represent the continuous inclined compression
fields in the web of the beam. The strut- and- tie model is proposed as a
generalization of the truss analogy applicable to a variety of design situations.
33
The truss analogy is a specialized form of the strut- and- tie model and can be
used exclusively in the design of B-regions. Other types of models which apply
to the wide range of D-regions occuring in the structure are then lumped under
the more general category of strut- and- tie- models.
2.2
model the entire structure with a strut and tie model. Rather, it is a more
convenient and common practice to first carry out a general structural analysis.
The general elastic analysis of linear structures results in determination of
external support reactions. Then, from equilibrium methods, sectional effects
(bending moment M8 , normal forces N, shear forces V, and torsional moments
Mr) can be determined at any desired section.
It is advantageous to subdivide the given structure into B-regions and Dreg ions. In order to roughly find the division lines between B-and D-regions, the
following procedure was proposed by Schlaich et al. [2]. It utilizes the well known
principle of Saint Venant which localizes the effect of concentrated forces as
shown in Fig. 2.1. This procedure is illustrated in the examples of Fig. 2.2.
1.
Replace the real structure (a) by the fictitious structure (b) which is loaded
in such a way that it complies with the Jakob Bernoulli hypothesis and
satisfies equilibrium with the sectional forces. Thus, (b) consists entirely
of one or several B-regions. It usually violates the actual boundary
conditions.
2.
34
Figure 2.1:
L. ,
....
Column
35
Beam
t " ! l l l l l l IIIII II
'1-~:
.~
(a)
(a)
~1'"'"'''""''1
CilD
(b)
(b)
F
cln'F;h
i
dh
.......
tdh-f
+<~
~}
'1"""'
dh
J-fl:Dt
<1'-h-
(c)
(c)
(d)
(d)
Figure 2.2
36
3.
Apply the principle of Saint-Venant to (c) bo that the local stresses may
be assumed negligible at a distance "d" ftom the equilibrating
forces, which is approximately equal to
;l.o
{i)
{2)
37
_j
(3)
(4)
Assume that the total area of zone 2 +zone 1 +zone 2 is the effective
D-reg ion
The discontinuity region boundaries may also be determined by the use
~J
B-region
1.5h
D-region h0
h zone 1
h'
B-region
h'
1.5 hcos B
Figure 2.3:
statical discontinuities
38
2.3
model for each application. For very unusual configurations as well as to ensure
good crack control behavior at service stress levels, it is recommended that the
model should be generally based on the principal stress pattern as determined
from an elastic analysis . For unusual cases such an elastic analysis with the
principal stress direction can be computed with an elasticfinite element program.
The directions of the struts and ties can be located at the center of gravity of the
corresponding stress fields. It usually makes it more convenient if the principal
stresses are converted to stresses parallel to the structural member borders {ax,
crY, 'txy,). A strut- and- tie- model based on such orthogonal elastic stresses chosen
parallel to the concrete surface generally leads to more economical straight
39
(2}
Subdivide the discontinuity zone into regions in such a way that the loads
on one side of the discontinuity zone are in unique regions with their
counterpart on the other side of the discontinuity zone. These regions are
the load paths connecting the opposite sides and tend to take the shortest
possible streamlined way across.
'
'
The load paths must be single lines and must not cross each other.
(3}
If the applied forces are not completely equilibrated with the obvious load
paths, then the resulting loads must follow a U-path as shown for 8-8 in
Fig. 2.4.
'__ j
(4}
Sketch all load paths {including possible U-paths} and replace them by
polygons made up of compression struts and tension ties.
(5}
Add further struts and ties as required for equilibrium at the nodes.
(6)
For very complicated cases, the finite element analysis results can also be
combined with the load path method.
___ _!
40
B C
o)
B C
I
II
I
II
)--\\
Tz/
I'
'\
eM-~
b)
.I
11.
II
Ii )
. f
/
u./
..~'! 'V
II
J__~
/\
tc
---
strut
tie
load P4t.h
anchorage length of the bar
Two models for the same case: (a) requiring oblique reinforcement;
(b) for orthogonal reinforcement.
Figure 2.4:
41
2.4
Strut Background
The struts transfer internal compression forces from node to node. In
general this transfer is via three-dimensional stress fields in the concrete. Due to
compatibility requirements the stress fields tend to spread out between the
nodes and fill the available space. This results in transverse tensile and
compressive stresses which must be considered in the evaluation of a strut's
strength and 'which may require provision of local reinforcement (see Fig. 2.5).
However, well tied compression reinforcement will also resist compressive
forces as long as it is restrained from buckling. While all concrete structures must
be built and reinforced in three dimensional space, it is usually sufficient to
determine reinforcement separately in two orthogonal planes. This leads to
consideration of two dimensional or planar struts.
Four typical configurations of two dimensional compression fields are
presented in Fig. 2.6. Generally it is safe to determine the strength of
compression struts using one of the four simplified types of stress fields shown
in Fig. 2.6 (a) to {d).
A prismatic strut as shown in Fig. 2.6 {a) is the simplest idealization of a
compressive stress field. The prism is uniform in geometry and has a constant
stress along its length. Prisms are generally used to model stress fields having
uniform parallel stress trajectories.
The fan shaped stress fields shown in Fig. 2.6 (b) are developed at points
of concentrated loading or at supports. Fig. 2.7(a)
J
42
.......
....
!":-
'
compression
Figure 2.5:
,_j
43
T
L
'
_I
(b) fan
(a) prism
4----8-~
'
' J
(c) bottle
I
Figure 2.6:
'1
44
T9
TS
5-Struts
T-Tension Ties
Bottom Chord
Figure 2.7(a):
'
Singular
Nodes
<a) Stress Field:
Figure 2.7(b):
<cl Corresponding
Reinforcement
45
reinforcement to properly develop the required tie forces. The confined core of
Fig. 2.6 {d) is a transversally reinforced core or prism with a special behavior. The
reinforcement can be spirals, closed stirrups or steel pipe. The reinforced core
develops under load a three dimensional state of stress, which is controlled by
the behavior and the form of the reinforcement and the transversal contraction
of the concrete. It is generally restricted to points of application of very large
magnitude forces or relatively small areas as when post-tensioned tendons are
anchored or when extremely large loads are applied by columns bearing on a
transfer girder.
46
High strength concrete requires extreme care in all steps of the production
process. It has become common practice to specify high strength concrete
strengths at 7, 28, 56, or 90 days [57, 58]. Economically it is important to know
at the outset of high strength concrete production specifically what strength one
needs and when one needs it. High strength concrete requires a very low water
to cement ratio (=0.25- 0.45}. Therefore, inaccurate estimation of the aggregates
water content, which affects thequantity of additional water added at batching,
can result in either balling of the concrete due to lack of mixing water or in too
high a slump. Mixing is critically important as well. For satisfactory performance
all the materials, especially admixtures, must be thoroughly mixed. Curing
becomes more critical in high strength concrete production and proper hydration
must be allowed to prevent shrinkage cracking.
The tensile strength is significantly linked to the curing conditions. The
following relations (concrete compressive strength between 731 0 and 10040 psi}
were given in Ref. [33] with a 10% coefficient of variation:
moist cured
dry cured
\_
47
GRAVEL
10
I High Strength
I
I
.J
-"'
-.,
&II
Cl.l
1\
I
I
~ t Medium Strength
h
11 Medium
I\
I\
I~ Strength
,,,,'
.....
(/)
4
,-
High-Strength
1
I
--~
''
2
3
4
56
Strain ( x 1000 in./ in.)
Figure 2.8:
'
,J
'',
~--.___.___~--~--~--~---
::
', 1
Normal
' srrength
'
~\Normal-Strength
70
4
5
3
2
Strain{ 1. IOOOin./inJ
48
It has been stated that the descending branch becomes almost a vertical line.
The ultimate strain at failure is lower than for moderate strength concretes. The
steeper stress-strain curve for high strength concrete means the modulus of
elasticity is higher. The following equation for the modulus of elasticity has been
proposed [59]:
0
O
0
c:
cs
Q.)
cU
Q)
c:
Q)
.!:!1
Q)
....Q)
CD
oe5
N
....
(I)
C'IS
a:
49
""
.....
.c:
c:
c: ~
C'IS
.5
;:::
::;::,
r=
....
0
0)
0
0
0
Q.)
oe5
....
Cl)
c: C)
=a
0
,...
C\1
(.)
C'IS
ttttt
0
0
0
0
,...
-....en
Q.
0
0
0
CX)
.c
....
0)
c
G)
....'-en
....
G)
0
0
0
<0
G)
'-
(.)
c
0
(.)
.. J
0
0
0
""
0
C\1
co
ci
.......
ci
<0
ci
ci
50
factor is that the redistribution of forces in the member due to the different ratios
of longitudinal to transverse reinforcement may cause the failure crack and the
compressive struts between them to be at an inclination otherthan the 45 degree
angle corresponding to initial diagonal tension cracking of the concrete. Thus the
strut may be crossing previously cracked concrete. Another important factor is
the need to select a very conservative value due to the undesirability of a failure
due to crushing oft he concrete in the web because of its brittle nature. In general,
the effective concrete strength available for use in the struts is chosen as some
portion of the concrete compressive strength f'c The effective strength f'c is the
product of an efficiency factor ve and the 28 day cylinder compressive strength.
The efficiency factor should take into account the following parameters:
multi axial state of stress
disturbances from cracks
disturbances from reinforcement
confining reinforcement
friction forces
aggregate interlock after cracking
dowel forces
time dependence
f'ce =Ve f'c
Various proposals for the efficiency ''ve" factor have been presented. They are
usually based on tests of continuous compression fields either in rather thin
web beams or rather thin shear panels although some seem to be based
largely on engineering judgement. Very little experimental verification exists
for effective compressive stress efficiency factors for use in model analysis or
for use in large panels where shear is not a major concern. Many of the
various proposals for the efficiency factor are summarized in this section. They
basically correspond to the product of a basic efficiency factor and a modifier
to make them applicable to thin webs although this distinction is not always
shown by the various authors. Another factor considered by some authors
was the fact that in the case of torsion the twisting of the beam induces an
additional compression stress into the diagonal. Lampert and Thurlimann [14]
stated that the increase in the diagonal compression test was due to a
distortional
effect
in
the
walls
of
the
cross
L_
51
~'
Figure 2.11: Forces acting on edge members of parabolic arches (from Ref. [23])
52
section. Through twisting, the originally plane walls of the section are distorted
to hyperbolic paraboloids (Fig. 2.1 0) limited by four straight edges. The distorted
wall then constitutes a hyperbolic parapoloid shell subjected to a uniform shear
flow. The entire shell when loaded in this fashion is subjected solely to pure shear
stresses of constant intensity (see Fig. 2.11 ). These edge shears require edge
members. In the case of truss models these edge members are provided by the
longitudinal chords which are thereby loaded axially. The additional compressive
stresses on the outer surface of the diagonal due to wall distortion must be added
to those obtained from the actual shear flow. As a result Thurlimann [ 60]
suggests that the maximum value of the compression strength in diagonal
compression struts used be approximately 2400 psi, corresponding to f'cof about
4800 psi. Thurlimann [60] on the basis of test evidence proposed that the
allowable efficiency factor for the compression stress be:
l~
53
_j
"I" longitudinal
t
10
Onsel of;,YieldOuj
longitudinal
a
tR ., Oisptocement Parometer
Er Yield Strain of Reinforcing Steel
6
4
0.5
:s
tan ex
2.0
Figure 2.13: Relationship between the mean crack strain and the strains in the
reinforcement for different angles of inclination of the diagonal strut
(from Ref. [61])
'--- .. )
54
not exact limits but give a general range for transitions of failure
mechanism.ThOrlimann noted that at "Ew I c;" values of about 5 the failure
mechanism begin to change, and either shear or flexural failures become
possible without both of the reinforcement types yielding. It is also shown that if
the angle of inclination is greater than 45 degrees, yielding of the stirrups
demands larger mean crack strains. Conversely, for angles less than 45 degrees,
yielding of the longitudinal reinforcement requires increasingly larger crack
openings.
A best fit curve of the form k (f'c )05 to approximate the equation proposed
by ThOrlimann resulted in the relation [23]):
34
ve
(psi]
(fc )0.5
Ramirez and Breen [23] suggested that the compressive stress in the
compression diagonals should be less then 30 (fc )0 5
A study by Hartmann, Breen and Kreger [33] investigating the shear
capacity of high strength prestressed concrete girders (fc=12000 psi) showed
that by using this concrete limit the experimental results were 1.72 times
(standard deviation = 0.26) higher then the expected results with the 45 degree
truss model (see Table 2.1 ).
Zimmerli [56]
function of varying strut angles (see Fig. 2.14) based on test results of beam
webs.
L .
l
'___;
'
55
1.0
--e-- efficiency factor
...
0.9
( ,)
0.8
>.
(,)
.!!
.2
0.7
Q)
0.6
0.5
'
10
. 20
30
40
50
angle [0 ]
60
70
80
Figure 2.14: Varying strut angle versus efficiency factor for concrete
compression strength
56
$cs.
Nielsen et al. [21] at the Technical University of Denmark applied a rigid
plastic model for the concrete based on the modified Coulomb failure criteria.
With the assumption of only plane stresses the model gives a square yield locus
with a compressive yield stress of f'c and zero tensile capacity. Based on
experimental results an effectiveness parameter for concrete strength in the
webs was suggested. Nielsen et al. [21] described the experimental facts by
means of an empirical formula of the type
725 s crc s
8700 psi
57
12 (h)
13 (Jl)
0.15 Jl + 0.58
Jl s; 4.5%
14 (a)
a/h s; 5.5
fs (cretlcro.2)
for prestressing
(jo
cro.2
(jeff
a
h
Jl
As
b
=
=
=
=
=
depth of beam
=
=
As/(bh)
longitudinal reinforcement
The analysis included 186 test results from normal reinforced rectangular beams
and 19 rectangular prestressed beams
statistical parameters of the ratio between the experimental and the calculated
ve
Table 2.2:
1.0
14.5%
prestressed
1.0
8.6%
58
For design Nielsen et al. [21] recommended the use of the conservative straight
line expression
ve
Limits are also placed on the angle of assumed strut inclination to prevent too
large a deviation from elastic behavior.
21.8 < -Pes< 45 :
Mitchell and Collins [t9] and the Canadian CSA-Standard A 23.3M84 [29] presented a more detailed method for determining the limiting stress
in compression struts based on results oftests on shear panels. (See Fig. 2.15.)
normal strain
L .
59
The principal strain may be determined based on Mohr's circle of strain using the
strut angle, principal compressive strain and the strain parallel to the beam axis.
The efficiency factor is related to the principal tensile strength along with the
cylinder compressive strength.
1
V=
e
Mitchell and Collins [61] assumed' that the principal compressive strain
(2) in the strut is generally- 0.002 and the principal strain direction is assumed
to coincide with the principal stress direction. The principal tensile strain in the tie
1 can be computed by compatibility as follows:
f'c 0.6
fc
fee
""'
state
1
partial safety factor= 1.667
0.6
:. .. )
ve
<1>
factor)
60
Table 2.3:
Structural member
ve [29] ve [62]
Truss nodes:
Joints bounded by compressive struts and bearing areas
0.85
0.85
0.75
0.65
0.6
0.50
0.85
0.50
L_
30
0.31
0.25
tPcs
45
0.55
0.45
r .
For buildings of normal importance the load factors for dead load and live load
are D = 1.25 (except that if dead load resists overturning, uplift or stress reversal,
then D= 0.85) and L= 1.5 respectively.
61
Relating to the Canadian CSA - Standard [29] the CEB-MC - Draft 1990 [64]
gives the following formula for plane stress fie Ids with closely spaced cracks and
no major geometrical disturbance:
ve
f'c
1
= - - - - - - - - :::; 1.0
(0.85 + 0.27 f\ 1c1)
In the CEB Code-Draft [64] , 0 = 1.0 in all cases and the load factors for dead and
live load are 1.35 and 1.5, respectively.
Major skew cracks are not likely, if the theory of elasticity is followed sufficiently
closely during modelling. This means that the angle between struts and ties
entering a singular node should not be too small. However, skew cracks may also
be left over from a previous loading case with different stress situations (creep,
shrinkage, temperature etc.)
The CEB proposed efficiency factors are related to specified safety factors
which are different from those in North America. The following equation will be
used in Europe to compute the effective concrete strength:
.J
fee
fee
=
=
f'c
c
=
=
1.5
ve
ve
62
Table 2.4 gives the efficiency factors proposed by Schlaich et al [28] and
CEB- MC 1990 [64]:
Table 2.4:
ve
1.0
0.8
0.6
0.4
l_
In the following subsections, some test results are analyzed with the strut- andtie-model in order to evaluate the efficiency factor for compression struts in
cracked webs.
63
2.4.1.1
Shear capacity can be analyzed with different models. A number of empirical and
conceptual models have been presented over the course of time. Given all of this
effort, however, a completely satisfactory solution has not been attained. The design
concepts of the ACI [3} and AASHTO [3]- recommendations and those in the CEBMC - Draft 1990 [64] have in principle the same structure. The general basis is
.J
vu
vc
vs
vp
q>
=
=
=
=
=
For reinforced concrete there are two equations for Vc under normal loading
conditions. One equation is [3]
vc
As
Mu
bw
d
=
=
=
=
=
{1.9 (f'c )05 + 2500 {A/{d bw) [Vu d I Mu} bw d ~ 3.5 (f'c )05 d bw
area of longitudinal reinforcement
factored moment at section
web width
distance from extreme compression fiber to centroid of longitudinal
tension reinforcement
vc
= 2 bw d {f'c )05
64
Vs
Av fy dIs
=
According to CEB [64] the concrete contribution is given as follow
=
Values for
Table 2.5:
'tRo
Values for
'tRo
f'c
'tRD
26.1
31.9
37.7
43.5
49.3
55.2
60.9
66.7
72.5
Kordina and Hegger [32] present another fo~mulation for the concrete contribution
in prestressed concrete girders.
=
=
65
An excellent approach is given by Vecchio and Collins [65] to predict the
response of reinforced concrete beams subjected to shear using the modified
compression field theory. Strain softening and tension stiffening effects are taken
into account in the theoretical model.
The strut- and-tie-model for shear design is based on some assumptions.
Yielding of both the longitudinal and transverse reinforcement is required. This
requires an upper limit on the diagonal concrete stresses to prevent crushing.
The reinforcement can only resist axial loads. The reinforcement is properly
detailed so that local crushing and bond failures are prevented. The angle of
inclination forthe compression diagonals differ as proposed by different authors.
25 ~ <~>cs ~ 65 (Ramirez, Breen [23])
The orientation of the diagonal compression strut and the width of the strut are
the most important factors for a strut-and-tie-model. Fig. 2.16 shows the strutand-tie-model based on the idea of a "design zone" for the ultimate behavior
under shear and bending proposed by Kaufmann and Ramirez [66]. A new
formulation of the compression width is given here and the proposed model is
compared with test results. To obtain this strut-and-tie-model, the beam is first
divided into design zones. A vertical tension tie is placed at the location of the
resultant force ofthe web reinforcement in each design zone. The tension chord
is located at the centroid of the flexural tension reinforcement and the compression chord is located at the centroid of the flexural compression block. Diagonal
concrete members are then placed to complete static equilibrium of the model.
66
REPRESENTS RESU..TANT
FORCE OF WEB REINFORCEMENT
WITHIN DESIGN ZONE
a
z =d-2
TENSION
CHORD
--~~
DESIGN ZONE
OESlGN
ZONE
<=V=O
Figure 2.16: Proposed strut-and-tie-model for shear behavior (from Ref. [66])
67
Table 2.6:
Ref.
f' c
Ca
cpa
$1
Ts
[67)
5.28
111.9
15.3
28.7
1B
20.5
0.435
[67)
7.44
161.2
15.3
28.7
18
20.5
0.44
[67]
7.44
142.2
15.3
28.7
18
20.5
0.39
[67]
8.1
170.8
15.3
28.7
18
40.9
0.43
[66)
8.34
287.4
21.8
34.8
24
39.3
0.67*
[67]
10.49
191.5
15.3
28.7
18
20.5
0.375
The following equation for the concrete efficiency factor in diagonal compression
struts is proposed (see also section 2.4. 1.2 confined concrete strength). As can be
seen in the next section the
68
same equation without the 0.6 reduction factor is used for the basic ef'riciency factor
for confined concrete strength and for the compression strength in the nodal zone.
The reduction factor 0.6 (in actuality a judgement factor "3/5") reflects the lower
effective
differentiation between the higher effective concrete strength in nodal zones and
isolated concrete struts as compared to more uniformly stressed webs is also made
by MacGregor [62] as reflected in Table 2.3.
proposing efficiency factors do not consider reductions for high strength concretes.
The statistical data 'fromthe comparison are given inTable 2.7 and shown in Fig.
2.17.
2.0
--
--o-- experimentltheory=1 .0
...0>-
-e
1.5
~mentltheory
""'-
..
II
....
1.0
I ~
II
;:::
II
c.
loC
0.5
II
0.0
eooo
5000
7000
8000
10000
11000
Figure 2.17: Comparison of test results with the theoretical approach of predicting
the diagonal compression strength
x, :Column 1
Dev.:
Mean:
Std.
11.136
,.272
Minimum:
Maximum:
.967
Table 2.7:
1.68
Std. Error:
Variance:
Coef. var.:
123.938
6.815
69
2.4.1.2
1.1
efficiency factor
threedimensional compressive strength
efficiency factor
Table 2.8:
vea
3.0
'l
1.1 f'c
-1.0/-1.0/0.
-1.0/-0.93/-0.18
-1.0 I -0.491-0.14
-1.0 I -0.50 /-0.25
-1.0/-0.261-0.09
-1.0/-0.25/-0.12
-1.01-0.16/-0.08
-1.0/-0.14/-0.14
-1.0 I -0.261 -0.09
-1.01-0.1/-0.05
strength ratios
1.1
==:6
3.5
=8
3.3
6.0
2.6
4.4
3.3
1.8
I
70
2~------------------------------~
--E
B
faa
E:J
en
en
51152=-1 J-1.
51/52=-1 J-.52
51/52=-1J-.22
C/)
4608
2716
8804
4480
Mean:
Std. Dev.:
X1: Column 1
Variance:
Std. Error:
11.264
1.114
,.033
Minimum:
Maximum:
1.123
1.498
1.013
Coef. Var.:'
Count:
18.991
Sum Souared:
15.1 S4
19.304
For the two- or three dimensional state of stress, a large number of theoretical
investigations have been carried out in recent years and various models have
been proposed to characterize the multi axial stress-strain behavior of concrete
. A brief review of some previous recommendations is given here. The Cauchy
model (nonlinear elastic) by Kotsovos [70], the hypoelastic material law by
Stankowski and Gerstle [71 ], the elasto-plasticconstitutive law by Han and Chen
[72] and the bounding surface model developed by Meschke et al [73] and by
Fardis, Ali be, Tassoulas [7 4] represent typical constitutive models for description
of the material behavior for multiaxially loaded concrete structures.
L~
71
_!
2.4.1.3
As
=
=
=
fy
db
1t
d 2/ 4
Ab
1t
db2J4
fe3
flat
f/ (d s)
d
Figure 2. 19: Typical geometrical data for confined core
72
total cross sectional area of the stirrups and ties (cross tie included)
If the square compression strut has no longitudinal reinforcement and the lateral
reinforcement consists of square ties, the effective confinement was found by
Fafitis and Shah [76] to be about 40% of the confinement pressure for a square
compression strut with longitudinal reinforcement.
Recent work work by Ahmad and Shah [77] has shown that spiral reinforcement
is less effective for compression struts of higher strength concrete. The authors
also found thatthe stress in the steel spiral at peak load for high-strength concrete
is often significantly less than the yield strength. These conclusions are consistent with results of experimental research at Cornell University. In the study by
Martinez, Nilson and Slate [78], an effective confinement stress "fs (I - s/d)" was
used in evaluating results, where "fs" is the actual stress in the spiral. The term
"(1 - s/d)" reflects the reduction in effectiveness of the spiral associated with
increasing spacing of the spiral wires.
fe3 =f' c (A/Ab)05 + 4 0
tact
lat
fs (I - s/d) 2 As I (d s)
compression load
Jl
fact
Iat
j __ _!
73
=
=
0.7 forties
0.75 for spirals
When the supporting surface is wider on all sides than the loaded area (Ab) , design bearing strength on the loaded area may be multiplied by (A I ~) 05 , but not
more than 2.
-..,J
.,J:l.
"11
u:r
c:
r-:>
1\)
om
0 X
2~J.- ~y
-""0
::I
<0
:::!.
en a>
~::J
ap.r
3 en
-
20
NCI&9112500~i1Jh Strn9~h---.
r;
:D~
(1)
(1)
...--.
.-en
en
-...Jen
(() r+
~~::::1
Axial
Stress,
150
151--f
Strntlh
Axial
100
llsi
MPa
c:
10 I--"
I Ha
en
.....
low
~H-IH
II-
II
1/ I
1/
Str~notll
--t:,;;
..,
P>
!SO
I BOO)
;:}:=:;;:~~---+,--._
NCial
::J
::J
Stress.
o~ 0 010 ..1
I
1
I
'"'''Ill
Axial
<0"
:::r
en
r+
""2.
QJ
.. I
(24<1
I
75
=:;;
50-65
length of shorter side of a rectangular plate or side of square plate
=
=
K
_j
76
For a=b=c and K=50 the increase of the bearing stresses is a function of the square
root of the concrete strength. The author proposed for design purpose a K equal to
50 (see Fig. 2.22). The results for block length equal to the plate dimensions are
shown in Fig. 2. 22.
In 1971, based upon further tests, Hawkins [84] recommended the following
formula for strip loading of concrete through rigid plates:
r
L
d/2
,,
Figure 2.21 (a}:
'
77
Niyogi [85] discussed the problems associated with the calculation of the
allowable stresses and the probable mechanics of failure. The primary parame'.J
ters were the geometry of the bearing plate related to the loaded surface and the
plate geometry. Square, rectangular and strip loadings were considered. He postulated the following formula:
fc3
Dimensions and definitions are shown in Fig. 2.21 {b). According to Niyogi, the
bearing strength decreased for increasing height and eccentricity of the load.
2a
2a'
Figure 2.21 (b): Variable definitions for geometry of the bearing and loaded plate
(from Ref. [85])
78
0
0
0
.q-
,...
'\
'"
N
,...
.,
1\
0
0
0
0
0
1\
.
,...
g
,...
co
,...
g
0
0
,...
J::
a
c
-e
CD
&..
en
G)
0
0
0
U)
0
0
c
0
u
0
'-
---
0
,...
eQ.n
'i
0
0
0
<0
0
0
0
0
0
0
<0
0
.q-
~
o
0
0
0
oC\1
...r
C\1
0
0
0
0
Figure 2.22: Bearing stresses versus concrete strength (from Ref. [83])
~J
79
-j
n,
fc3 [1 - (1 - elec3)A]
fc3 exp[ -k (e- ec3) 115]
=
=
Ec ec31 fc3
33 w1.S (f'c)o.s
=
=
fore~
ec3
fore;;::: ec3
so
The analytically predicted values of peak stress compared to the test results (with
concrete strength of the specimens from about 3000 psi to about 10,000 psi and
the confinement pressure from about 250 psi to about 3000 psi) gave differences
from 3.5 to 26.5%.
Schlaich et. al. [2] propose the following equation to compute the confined
strength for spiral confining reinforcement:
For square compression struts and square confinement reinforcement the lateral
pressure can be computed by reduCing the equivalent circular compression
strength by 50%
Roberts [89] tested local anchorage zone specimens with spiral confining reinforcement. Test results of five different authors were compared with theoretical
approaches to determine the best fit function. For the basic concrete efficiency
factor the same term was used as shown in 2.4.1 .1 for concrete strength in compression diagonals. The following approach is used for the comparison
81
author
specimens
experiment/theory
mean
standard deviation
Roberts [89]:
28
0.95
0.15
29
1.05
0.06
Niyogi [85]:
39
1.07
0.19
Muguruma [88]:
25
1.38
0.22
Martinez [78]
11
1.31
0.09
Total
122
1.125
0.24
Different sensitivity analyses have shown that the reduction factor (1 - s/d) fort he
confinement strength has a significant infJuence. In the following approaches the
squared reduction factor (1 - s/d)2 is used for the comparison. For practical
application the reduction factor forces the designer to use smaller spacings for
confinement reinforcement.
82
The term "(1 - s/d)" reflects the reduction in effectiveness of spiral associated with
increasing spacing of the spiral wires. For a better correlation with test data the
second term in the equation was changed. The following approaches were used:
f63 = [0.5 + 15/(f'c )05 ] fc (A/Ab)o.s + 4.0 {Acore/ Ab) f181 {1- s/d)2
Comparison with the test results {see Fig. 2.23 and Table 2.12) shows that the
proposed equation with the effective confinement strength is a generally conservative and safe approach. The 95 percent limits (X- 2cr) would be 0.65 which is
also the minimum actual test result.
Table 2.12:
Mean:
S.d. Oev.:
X1: CoNmn1
Std. Error:
Varianee:
11.124
1238
1.022
Minimum:
Maximum:
.054
2.128
l.o57
Coat. Var.:
Count
121.184
1120
134.851
"
the statistical
mean of the comparison is 1.05 (see Table 2.13). The 95 percent limits become
0.61 with the minimum actual test result 0.62. Either of the.se efficiency factors
could be used in practice.
Table 2.13:
ve = 0.5 + 20/(f'c )0 5
0
0
83
~------~------,---------------~-------ro
~
....
0
0
C\1
....
0
0
0
0
.... ......
Q.
-...
-...
.c
0)
80 cG)
a:>
.....
G)
II
~~
G)
(,)
-Q)
0
0
0
Q)
.r:. .r:.
cQ) cQ)
U)
(,)
E E
Q)
Q)
a. a.
X
Q)
0
0
0
Q)
tt
ll)
C\i
C\i
0
0
0
&0
oN
0
experiment I theory
Figure 2.23: Comparison of test results with theoretical approaches for
predicting confined concrete strength
84
2.5
Tie Background
The
,_
'
l.
85
86
-i ii
800
6.0(fc)E.5
1.7(fc)E.67
0.44(fc)E.6
o.osrc
c.
=
C)
-..
cG)
600
.!
iii
c
400
.!
.!Gl
loo
c
0
2000
3000
4000
5000
6000
7000
8000
9000
Figure 2.24: Various approaches for the tensile strength of concrete [91]
1.0
-#.
-..
.s::
0.8
C)
0.6
0.4
experiment
1/(1000W+,)
Gl
en
.!
iii
cGl
1-
0.2
0.0
0.00
0.01
0.02
0.03
Figure 2.25: Crack width versus tensile strength reduction factor (from Ref. [53])
L_
87
f' c
0.25
1.0
'_j
88
Intensive research work has been done to develop procedures for the calculation of the flexural strength of prestressed concrete structures. Also a large
number of prestressed concrete specimens have been tested to determine their
strength in resisting shear, or combined moment and shear, with or without web
reinforcement [91].
Prestressed concrete members pose a somewhat different challenge in
detailing. The computation ofthe ultimate strength of such members is generally
very similar to that of nonprestressed members when appropriate assumptions
and calculations are made to determine the stress level in the tendons at failure.
This stress level depends principally on whether the tendons have been effectively bonded to the concrete. Required quantities of non-prestressed reinforcement to resist anchorage force concentrations, to distribute support reactions,
and to counteract similar concentrated load effects can be effectively determined
using strut- and- tie models for these D or discontinuity zones. Use of such
models actually gives superior understanding and very good guidance for the
special reinforcement required. However, while they can be used if desired, strutand- tie models applied to the B zones tend to be time-consuming and somewhat
unwieldy in application.
Since strut- and- tie modelling is a plasticity approach, it is quite inefficient
and artifical to use if for checking elastic stresses such as allowable stresses at
service load levels. Such checks are important in prestressed concrete members
in which it may be desirable to prevent or severely limit crack formation at the
service load limit state. It is possible to make such calculations using a type of
strut- and tie model forB zones, but it does not appear very efficient to do so. In
order to illustrate the nature of the computations, a type of model adaptable to
allowable stress calculations will be shown in this section and used in Example
4.6. Examination of the actual calculations will indicate the general impracticality
89
90
year [92]. The loss of prestress due to shrinkage is the product of the effective
shrinkage "esh" and the modulus of elasticity of prestressing steel.
esh
RH
V/S
= 8.2 r1
os
relative humidity
volume-to-surface ratio
L ....
L~
L_
o..
91
'*w~r---+-~..-_-_-_-.-_-_-_...:::.:.._-_-_-_-_-..J-+--i-c-_-___,::j ~
(a)
r----- ----.,
'~B
s ......1.....,..
!.,!pi-losse
... 1 _ __;.;.........:;.;.
___
.,....,!-'-.
L---------.J
P~e<P,
I. Prestress alone
Tens.
Tens..
-~[
-~
].p"
c ....
on A,
Comp.
~
l+
~f ~
Md+ M! on It
Tens.
cY
~
'"'"''yr.;'
I.~
'~
~'-."~w
'l
=
Tens. c
6...;;:
(c)
(b~
Ultimate comp.
Cracked
(d)
(a)
{b)
(c)
(d)
92
Creep is assumed to occur with the superimposed permanent dead load added
to the member after it has been prestressed. Loss of prestress due to creep is
computed for bonded members from the following expression:
pier
Kcr
Kcr
Ec
fcir
fcir
e
I
A
Mw
=
=
=
=
=
=
=
=
=
r~
'
l_
r-
'
L_
a
ll
=
=
=
lfw
=p
lfw,x
e (Jla + k I)
L~
93
The proposed friction- "Jl" and wobble coefficient "a" are shown in Table 2.14
[31 ]:
Table 2.14:
Type of strand
Friction coefficient
Wobble coefficient
)l [/]
a [klft]
0.15-0.25*
0.0002
0.15
0.0002
0.23
0.0002
0.25*
0.00
0.0002
Figure 2.28: Frictional loss along circular curve {from Ref. [92]
'
94
Loss of prestress due to steel relaxation overt he time interval t1 tot may be estimated
as follows for ordinary stress relieved strand::
P15 = [log ( 24 t - log 24 t 1 ) /1 0]
fs/ {0.85 fy) - 0.55
~0
As proposed in [31] for low relaxation steel a different expression can be used:
P15 = [log ( 24 t - log 24 t1 ) /45]
The value for t 1at the time of anchorage of the prestressing steel shall be taken as
1/24 of a day so that log t1 at this times equals zero.
pretensioning [%]
post-tensioning [%]
r
L__
elastic shortening
creep of concrete
shrinkage of concrete
steel relaxation
Total loss
25
20
L __
95
After an accurate estimate of the prestressed forces is obtained, the
spread of the highly concentrated forces into the member must be approximated
in the strut- and- tie models. Based on comprehensive finite element analyses
by Burdet (42) and comparison with experimental values by Sanders (43) a
conservative values of the diffusion angle was chosen. The proposed compression strut diffusion angle is 12. It is the first term from the more comprehensive
expression for the diffusion angle for a compression strut given in Section 3.4.
In post-tensioned concrete the point of application of the major or initial
force is relatively clear. Except for frictional and radial forces in curved tendons,
the post-tensioning forces are applied at the anchorages. However, in pretensioned concrete, the initial prestressing forces are distributed over longer
lengths. With both bonded post-tensioned tendons and bonded pretensioned
tendons, subsequent stress changes can be induced by flexural actions.
The development length of the prestressing strands is another important
_j
factor. Two types of bond strength must be considered. The first of these is
referred to as "transfer bond stress" and has the function of transferring the force
in a pre-tensioned tendon to the concrete. Transfer bond stresses come into
existence when the prestressing force in the tendons is transferred from the
prestressing beds to the concrete section. The second type of bond is termed
"flexural bond stress" and comes into existence in pre-tensioned and bonded
post-tensioned members when the members are subjected to external loads
[93]. After cracking, the increase in steel stress above effective prestress
develops flexural bond stress between the steel and the concrete. Flexural bond
stress does not exist in unbonded, post-tensioned members. Transfer bond
involves the Hoyer effect. When a prestresing tendon is stressed, the elongation
of the tendon is accompanied by a reduction in the diameter due to Poissons'
effect [94]. Hoyer [95] pointed out that, on release of the wire from its temporary
anchorage on the prestressing bed, the end of the wire swells as a result of the
recovery of the lateral contraction and develops a wedge effect because the
prestressing force must diminish to zero at the end of the wire. The stress in the
wire is zero at the extreme end and is at a maximum value at some distance called
the "transmission length" from the end of the member (see Fig. 2.29). The length
over which the prestress transfer bond exists is termed the prestress transfer
length, and depends mainly on the amount of prestress, surface condition of the
strand and the concrete strength. Three factors which contribute to bond
performance are adhesion betweeen concrete and steel, friction between
concrete and steel and mechanical resistance [96]. Libby [93] pointed out that
under normal conditions, the transmission length for clean seven-wire strands
can be assumed to be equal to 50 times the diameter of the strand. The
transmission length of tendons can be expected to increase from 5to 20% within
one year after release as a result of relaxation.
96
fransri!J":t
/e/J$/f:1
!J()IIf/sfress be/ween
wire fJI1daJ~~crd!
fJontllengfb ii
11: I
\\ll
Figure 2.29: Bond stress distribution at the end of a bond anchorage of a pretensioned wire [94]
97
The reserve capacity "TP11 of the prestressing steel which is still available for use
as a tensile chord after prestressing is equal to its yield force (under ultimate load
the prestressing steel is strained beyond its yield strength) minus the prestressing force which is applied to the member as a load
TP=AP fPY - P
p
___ )
fpi
fpy
AP
=
=
=
=
A major problem in the calculation of "elastic stresses~~ using the strut- and- tie
models is that the fully plastic strut- and- tie model does not adequately represent
98
the compatibility effects so important at the service load state. A good example
is the effect of tendon eccentricity. As shown in Fig. 2.30(c), the elastic stress
distribution can result in tension on the tob fiber. If the fully plastic strut- and- tie
model shown in Fig. 2.30(d) is used, the free rotation of the joints possible in the
assumed fully plastic members results in concentration of stresses in the lower
chord only and no forces in the upper chord, verticals or diagonals.
From
required on the top fiber area and along the support face jf concrete tensile
stresses are not considered to adequately carry the tensile force. This tensile
force, T 1
is the force which must be applied to any strut- and- tie model used
to represent the prestressing effect (see TP in Fig. 2.30(b)). In Fig. 2.30(b), the
chord forces shown are nominal forces. The load factors as well as the material
reduction factors have to be taken into account in design. In post-tensioned
members, if no bond is provided after prestressing, the prestressing steel cannot
be considered as reinforcement. The tendon force is applied as an applied force
or the tendon is considered as a constant force tie. Figure 2.31 shows the
proposed strut-and-tie-models for a prestressed concrete member with curved
or harped tendons.
99
,- -,
.,.Cp1
cpp1
"""' """'
....."""' .....
..... .....
I
I
"""'I
'' '
I
f."""'
Tp1
Cp1....,
-cp2'
"""'
"""' """'
- - - """'-
"""' """'
"""' """'
"""' """'
..... .....
..... .....
..... .....
..... .....
..... .....
..... .....
~I
=50 ~b
(a)
r
h
L
Tp :: wt b (PI At+ P
e \t I It)
_)
Cp
(b)
;-
web( PI Ac- P e
Yc I
~)
Overall Model
100
(c)
(d)
(e)
Strut-and-Tie
101
~~~~~~~~w:c
e =f{x); z = f(x)
p~
Cr1
~p
r
h
Tentlon chord
Vertical chord
=!
TY {load)
Cr (prettreulng)
102
Should it be desired to use strut- and- tie models in 'B' regions it can be done, if
inefficiently. This section presents a summary of the design steps involved in
such use ofthe strut- and- tie approach for prestressed concrete members. Table
2.16 show the analysis procedures for prestressed beams using the strut and tie
model. The symbols are related to Fig. 2.30 and Fig. 2.31 and are defined as
follows:
Table 2.16:
Design steps for 'B' regions of prestressed beams using the strutand-tie-model
step
load situation
selected model
strength limits
wires cut
a= 12
(1) Pretensioned
(1) Bonded or
Unbonded
PostTensioned
(2) Pretensioned
Bonded or
Unbonded
PostTensioned
immediately
after prestressing
4>
=12+3/(a/h)
anchorage zone
analysis (see 4.2.6)
'
. r-',
'l_-
Prestress+
Dead Load
compression chord
[T(prestr.)- C(load)-11
0 (tension)
I [T(Ioad] - C(prestr.)]
:!) b W1 v. f'c
or service state values
L_
tension chord
(3) Pretensioned
Bonded or
Unbonded
PostTensioned
Prestress+
Dead Load
+Live Load
compression chord
tension chord
verticals
r-\ ___
,----
(4) Pr:etensioned
Bonded PostTensioned
ultimate
compression chord
IC{Ioad}sbwc v. fc
tension chord
IT (load} S T + P
l
(5} Unbonded
PostTensioned
ultimate
compression chord
I C (load) S b we
tension chord
IT(Ioad)ST+P
v. fc
L
103
Load situations:
II
dead load
Ill
live load
v*
e
TP + Ts
TP
AP fPY - P
Ts
fpy
=
=
=
As fsy
yield strength of the pretressed steel
AP
fsy
=
=
=
AP
For curved tendons the radial compression component "Cr'' can be computed from
the deviations between the tangents to the curve (Fig. 2.32). The section length
should be chosen according to the spacing of the vertical tension members (stirrups
or lumped stirrups).
P as a tension force
104
cr
where
cr
sl
=
=
If the prestressed tendon is following a parabolic curve a uniform load over the
length of the span can be computed (see Fig. 2.33):
wb
8 Ph I (L)2
prestressed force
=
=
sag of parabola
length of span
-- _
-'
tendon
_._,,.-----.......Parabolic
... p
-h
L
Uniform load
' '
""
Concrete as freebody
Fig. 2.33:
The width of the compression - and tension chord can be found according to Fig.
2.34. The half of the width is the distance from the center line of the chords to the
outside fiber of the member.
pression chor
Fig. 2.34:
tension chord
wt
___!_
L-.__1
105
2.6
Node Background
The nodes of the strut- and- tie- model represent the locations of change
of direction of internal forces, which in the structure occurs over a certain length
and width in the node region. The intersecting strut- and- tie forces have to be
linked together and balanced in equilibrium in the node region.
If one of the struts or ties represents a concentrated stress field (e. g. near a single
load, a support or concentrated reinforcement) the deviation of forces tends to
be locally concentrated and the node region is relatively small. These kind of
nodes are called "singular nodes" and have to be dimensioned with special
care. The special studies about CCT (Compression-Compression-Tension)- and
CTT (Compression-Tension-Tension)- nodes given in Ref. [38, 39] and summarized in Report 1127-1 provide information upon which design recommendations
are based.
Splices or joints of overlapping reinforcement are the special but frequent case
of nodes occuring between two ties or reinforced struts for which specific rules
based on comprehensive tests have to be used.
Where wide concrete stress fields join each other, the node region extends over
a considerable length of struts and ties. Such "smeared nodes" need not
be
checked for safety, if the same D-reg ion contains a singular node.
Numerous possibilities exist for detailing nodes. In all cases, the flow of forces
can be visualized by strut- and- tie- models with singular nodes at the deformations of the bar. Bond is in fact a load transfer mechanism involving both
compressive and tensile stresses.
Singular and smeared nodes may be grouped into subsets relating to the
type of elements which they join. Four different kind of nodes can be worked out
from a strut- and- tie- model (see Fig. 2.35).
CCC: Compression- Compression- Compression
CCT: Compression- Compression- Tension
CTT: Compression- Tension- Tension
TTT:
106
(a) CCCNode
(c)
CTTNode
(b) CCTNode
(d) TTTNode
Evaluation of the nodal regions includes checking the nodal boundary stresses
and determining reinforcement development requirements for nodes which contain tension ties. Each of these steps requires the determination of the physical
boundaries of the node. The dimensioning of nodes is largely determined by two
l_
constraints:
All the lines of actions of struts and ties as well as any external forces must
concide
The widths and relative angle of the struts and ties constrain the nodal
geometry
L_
107
If the nodal geometry can be varied it should be chosen to minimize the stresses
in the nodal region. This is accomplished by selecting a geometry in which the
stresses along the border of the node do not exceed the limiting value of the
effective concrete strength (fce=V6 f'c). In order to get a state of planar hydrostatic
stress, the geometry should be selected so that the stresses on all the node faces
are equal. Both principal stresses within the nodal region would then equal the
stress at the boundary of the node [27, 28].
For a CCC-node under a hydrostatic stress state the strut forces are proportional to their width and the sides of the node are perpendicular to the axis of
each of the struts. It should be recognized that the geometry of the model may
not allow for equilization of the boundary stresses. Such a situation is shown
in Fig. 2.35a. Following Schlaich and Schafer [2], this stress state is tolerable if
the maximum ratio of stresses between any two sides does not exceed 2.0. In
order to get an hydrostatic state of stress, the geometry of the node can be
changed as shown in Fig. 2.36b. The intersection of the strut centerlines actually
lies outside the nodal region in this case. Bottle-shaped struts are often used
where one of the nodal boundaries is fixed as in the case of a node adjacent to
a bearing plate. A reduction of the width of the struts is required to produce a
hydrostatic state of stress. In this case a more convenient approach proposed by
Schlaich and Schafer [2] can be used to check the concrete strength in the nodal
zone (see Fig. 2.37):
ao
O'b
()co
=
=
=
108
(a)
(b)
_)
109
II
Ill~
II III
W2
/
o1
ao =
(hydrostatic stress)
110
For design purposes some general rule has to be adopted to check the stresses
in a CCC-node. The stresses from the struts without bearing plates must be
checked with dimensions relating to the bearing plates. It is very useful to
subdivide the node under the bearing plate into two sub-nodes as shown in Fig.
2.38. A key assumption is the distance from the centerline of action of the force
to the sub-nodes.The distance proposed is the quarter width of the bearing plate
(a/4). By studying the strut- and- tie- model shown it can be seen that for CCCnodes with only one bearing plate, the angle of inclination of the struts relative to
the plate becomes an important factor. In order to conform with test resultsof
post-tensioned anchorages, the CCC-nodes are best split into two parallel subnodes [42, 43] as shown in Fig. 2.38.The strut- and- tie- pattern depends on the
bearing plate width, a, because for increasing width a decrease in the required
tie force "T" must follow. The design concrete efficiency factor v recommended
in Section 2.8 is a lower bound solution and compared to test results gives a safe
approach. The stresses underthe bearing plate should recognize the degree of
local confinement and be checked utilizing the proposed expressions of Roberts
given in Section 2.4.
cc1
a/2
;)
oE
a/4
a/2
~
a/4
II
I
Figure 2.38:
CCC-node
111
The two approaches to compute the stresses in a CCC- node are compared in
Fig. 2.39. It can be seen that for the "hydrostatic" stress solution the width of the
ho1izontal compression strut is dependent on the compression angle and on the
width of the bearing plate. For the "quarter width of the bearing plate" solution the
horizontal compression strut width is dependent only on the bearing plate width.
The comparison is made for equal compression angles with the following
equations:
- hydrostatic stress solution:
ao =a tan
- quarter width of the bearing plate solution:
a
ao
<!>
<!>
+ tan
=
=
3.0
:e
--111-
--.--
lJ
.!
Ill
2.5
"C
Ill
Cll
2.0
iS.
01
c
"C::
1.5
.e
e1:1.
E
1.0
u
ii
0.5
2
"C
0
.s::.
0.0
+-....;r-...,--.----r-.....--.,--....-__,.-----t
0
20
40
60
80
100
Fig. 2.39:
<!>3 )
112
For the CCT-nodes, two different reinforcing details can occur. Theoretically, but
rarely occurring in practice, the anchorage of reinforcement can be developed by
anchoring the tie forces from behind with an anchor plate (Fig. 2.40). The usual
and more practical case is anchorage by providing sufficient development length
behind the node as shown in Fig. 2.41. When using an anchor plate, the determination of the node geometry is clear. Special care should be taken to provide
adequate bending strength and stiffness in the anchor plate (plate bending
results in higher bursting forces) and to provide a proper connection with the tie.
A smooth surface for the tie where it crosses the node is theoretically better than
a good bond quality because strain compatibility with the bonded bar will tend
to crack the concrete. Bearing plate anchorage of tie forces usually means
diversion of compression fields. The compression stresses of the stress fields
concentrate on the steel plate's surface, if the tie is developed in this way. The
curvature of a deviated compression field is largest at the origin immediately
adjacent to the bearing plate.
In the more usual case of reinforcing bars directly anchored without
plates, either straight bars, hooks or loop anchorages may be used. Loop anchorages with confining direct pressure as from a bearing or direct load point are
preferred. Hooks shall preferably be placed to have confining pressure transverse to the hook plane. Sufficient anchorage lengths have to be provided within
as well as behind the node, if necessary. Anchorage begins where the compression struts (see Fig. 2.41) meet the surface of the bar. The bars should extend
to the other end of the node region in order to engage the outermost fiber of the
deviated compression strut.
For the effective widths of the struts- and- ties different proposals can be
found in the literature. The equilvalent concrete area approach [23, 24, 56] (see
Fig. 2.42) describes the width as follow:
113
Bearing plate
wcl
wT3
_L
Beartng plate
wc1
Figure 2.4 i: Anchorage detail for CCT-node with directly anchored bars
114
Wh
(1-2 W)h
Tension T~e
~
Wh
l_
v
Figure 2.42: Equivalent concrete area approach to define the tie width (from
Ref.[56])
Further proposals in the CEBMC - Draft 1990 [64] SI.Jggest that dimensions of
r ,
115
w3= 0.21d
L
1----;.o.region ~
~ID-region ~
cases anchorage failure was obtained. The approach shown in Fig. 2.41 , to
define the geometry of CCT-nodes for anchored reinforcing bars anchored by
development length behind the node, is based on the test results from Bouadi
[39]. In his specimens, the compressive forces and the tensile force in the reinforcement bar were increased simultaneously. All specimens experienced postyield failures including strut crushing, cover splitting, and gross slippage of reinforcement. In order to find the concrete strength efficiency factor for the CCTnode the specimens with concrete failure are compared in Fig. 2.45. The
_)
116
2340
Ba:s
I Layout
Bars
Layout
Full
u=o
247D
rull
LHT
2490
Half
LHO
2600
Half
LFTR
2610
Full
Plate
L_ __
4860
Fun
HFOSS
5005
Full
6-#7
HFO-HS
5015
Full
'6-#7
Hr=T
H~OSL I
Hi='OHL
5025
50.25
Full
16-#7
Full
6#7
Specimen
I&:J
'
L.
[Q
' . ~ :.
' .
I~
failure load
LFT
2360 psi
260 kips
14*12; 12*12
crushing of concrete
LFO
2360 psi
260 kips
14*12; 12*12
bond failure
LHT
2360psi
240 kips
9*12; 6*12
crushing of concrete
LHO
2360 psi
240 kips
9*12; 6*12
crushing of concrete
LFT
2360 psi
345 kips
14*12; 12*12
crushing of concrete
HFT
4860 psi
538 kips
14*12; 12*12
HFO-SS
4860 psi
450 kips
14*12; 12*12
HFO-HS
4860 psi
415 kips
14*12; 12*12
HFO-SL
4860psi
470 kips
14*12; 12*12
HFO-HL
4860 psi
433 kips
14*12; 12*12
Figure 2.44: General information about the tested CCT-nodes (from Ref. [131])
L.
L_
117
A CCT-node can be analysed by checking the concrete strength after finding the
geometry based on the approaches shown in Fig. 2.41. In order to optimize the
CCT-node both stresses at the C1 and C2 faces should be the same (hydrostatic
stress). The stress at the strut C2 face depends on the strut width "wc1 the tie width
11
WT3.. and the angle .. <Pes... Fig. 2.46 shows the geometric inter-relation of these
factors with various strut angles .. <Pes... The relation may be used for dimensioning
the width of the strut or to change the strut angle. The best way to design a CCTnode is to strive for hydrostatic stress (cr1 =cr2 =cr3 =1.0 where cri is the stress on
node side i) which leads to an optimal efficiency. The following equation can be
used to find the optimal solution (see Fig. 2.46):
=
1.6
experiment/theory
experiment/theory=1.0
-Q--
1.4
>....
0
-(1,)
c(1,)
E
1.0
(1,)
0.8
-....
c.
)(
.:..
.I:
1.2
(1,)
0.6
0.4
Al
B2
A2
A1-R
specimen
Figure 2.45: Comparison of test results with the concrete efficiency factor of
ve = 0.8
118
x,: Column 1
Mean:
Std. Dev.:
,1.173
1.138
Minimum:
Maximum:
Std. Error.
Variance:
Coef. Vat.:
1,.773
1.338
Table 2.17:
,....
2.0
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
W3=1.0w2
W3=0.5w2
W3=2.0w2
0.0
0
20
40
60
80
119
120
,_.-
t /
.
Design
Considerations
C1TNode-Strut-and-Tte Model
Strength of Members
Adequacy of Connections
Figure 2.47: Comparison of design rationale used for nodal region of strut- andtie- model and joint of steel truss (from Ref. [7])
121
Tension Tie
Itt t
'
LwT2
__!
wc1
_.J
4'cs
Figure 2.48: Geometrical approach to define the strut width for CIT-node
' .J
122
u.o
'7.5.
r_
Specimen
f~
.. 2Bday
fallure load
strut width
""""'
V' 'f"
I
'
lEnd
Vlewl
rl
anchorage detail;
confining transver:se
,--,
'
failure mode
reinforcement {yes-no)
HFSFI~A
7010 psi
121.4klps
10.6,. 12
noneH:aP.ofse(Up
HFSR-6
5780 psi
137.5 klps
10.6 "12
non~.ofsetUP
HHSFI
5780psi
139kips
4"12
none-cap. of setuP
HFSB
5780 psi
'138.1 k~s
10.6. 12
gross sJJp..trans.
HFNC
5780 psi
132.5 kips
10.6. 12
180 hook; no
cover splitting
LFSR
. 3720psl
117.4 kips
10.6. 12
, 80 hOok; yes
development-trans
LHSR
3720 psi
130.2 k4Js
4 12
strut crushing
LFNC
3720 psi
117.8 kips
10.6 11 12
, 80 hOOk; no
cover splitting
II..FAC
3920 psi
165.4 kips
10.6. 12
development-long.
Table 2.18: General information about the tested CTT-nodes (from Ref. [33J)
123
.J
The reinforcing
anchorage detail was primarily responsible for limiting the ultimate load. However, for design purposes the actual efficiency factor for the concrete compressive strength is of interest. Only one specimen (LHSR: f'c = 3720 psi) failed by
concrete crushing. The bearing plate stress was 3836 psi. By using a concrete
efficiency factor of 0.8 and by taking into account the smaller bearing plate width
(4") compared to the compression strut width (6.37") the experiment I theoryratio can be computed:
Cl~
3836 psi
1841311(12*4)
:::;
0.8 * 3720
3836 psi
(6.3714) 05
383613754 =
3754 psi
1.02
For this specimen with a concrete strength of 3720 psi a concrete efficiency factor
of 0.8 could be safely used. The efficiency factors for CCT-and CTT-nodes must
produce members in which the critical section will exhibit ductile behavior under
extreme overload. This is done by ensuring that actual failure would occur only
after the reinforcement yields. In order to guaranty ductile behavior, it is
necessary to place a limit on the failure state stress levels in the concrete.
When designing a CTT-node the reinforcement in both ties should yield at the
same time. In order to find the optimum strut angle for a given reinforcement pattern,
the following approach can be used (see Fig. 2.49). It is a geometrical approach and
is based on the compression strut width. Since the compression strut width "w1" is
dependent on the tension tie widths 11 W2" and "w3" shown in Fig. 2.48, the optimal
_j
concrete efficiency for a CTT-node is given by the angle with the largest compression
strut width "w1". The compression strut width can be computed:
~J
w1
Fig. 2.49 shows the compression strut width "w1 .. for various strut angle" <Pes .. and
three different tension tie width ratios.
124
3:
...
3:
...-
.;::::
n:l
..c:
"CC
"i
...
.=
..
c:n
1.0
0.8
0.6
;;;
O.A
C1)
0.2
c:n
c.
E
0
0.0
(,)
20
AO
W3 == l.Ow2
W3 =0.5w2
W3 = 1.5w2
60
80
2.6.4 TIT-nodes
In those rare cases where the tensile strength is used as a tension tie, some
global understanding about tensile strength has to be formulated.
Although is it difficult to develop design criteria fort he case of concrete tensile
ties, it would be even worse to maintain the formalistic view that the tensile strength
of concrete cannot and therefore must not be utilized. Tracing the flow of forces in
actual structures, to be gap free and consistent with strut- and- tie- models will
sometimes show that equilibrium can only be satisfied if ties or tensile force can be
accepted in places where, for practical reasons, reinforcement cannot be provided
and the tensile strength of concrete is implicitly utilized (see Fig. 2.50).
125
Tensile force
Figure:
Tensile force
may
would inhibit the development of concrete ties. For most practical detailing
problems, concrete tensile strength may be ignored. For those cases where the
tensile strength is needed, a value of
can be used.
126
lfthe tensile forces are transferred with reinforcing bars, the anchorage requirements became important. Anchorage is achieved by providing proper
devolopment length orin special circumstances by attaching the reinforcement
to bearing plates or other fixed components. The key to determining anchorage
requirements is selecting the point at which the reinforcement must be fully
developed. When the ties at a node are to be fully developed, a conservative
approach is to assume that the development length for each layer of tie
reinforcement is assumed to begin at the intersection point of the different ties
with the confined joint boundaries (see Fig. 2.51 ).
Conservative Location of
Critical Section for Computing
Developmertt Length
~
T
,_J
127
Positive anchorage details are those which do not rely appreciably upon
bond stresses to resist the applied tensile force and include end plates and
continuous reinforcing details. The positive anchorage detail must be designed
so that the tie force is distributed over a sufficient area to prevent the node zone
from being overstressed in compression. End plates and continuous reinforcement details are attractive from a design standpoint because they are fafrly easy
to evaluate.
Development anchorage details are those which are anchored with bent
bars (hooks). bond strength or a combination of both. Development anchorage
details are normally more economical, easier to fabricate and to place in the
formwork. The disadvantage of the development anchorage details is thefonger
required anchorage length.
128
anchorages
,---.,'
continuous looped reinforcement "positive anchorage details." The positive anchorage must be designed so that the compression resulting from the tie force
is distributed over a sufficient area to prevent the node from being overstressed.
r,
L.
However, such positive anchorages are not necessarily required nor are they
always desirable or practical construction alternatives for anchoring tensile ties.
Except for small diameter reinforcement, positive anchorage details are more
expensive and more difficult to construct than standard details such as straight
bars or hooks. Where the transfer of strut- and tie- forces is felt to be so abrupt
L_)
designer must check the development length requirements of the tensile tie
reinforcement. Sufficient development length should prevent splitting of the concrete cover and the resulting anchorage failure. Tepfers [97] suggested the
following approach to prevent splitting of the concrete cover for short anchorages
without transverse reinforcement. The cracking resistance, f'be , lies between the
l._J
f' be
f' be
=
=
f' be
cracking resistance
f' be
tensile strength
concrete cover
db
(0
average angle between the transverse cracks and the axis of the
bar = 45
r~
'-~
129
Bar development length '\" is the necessary embedment to assure that a bar
can be stressed to its yield point with some reserve to ensure membertoughness
under specific containment conditions. The necessary length is a function of a
number of variables, mainly of the bond strength and confinement from both
concrete cover and transverse reinforcement.. A great amount of research work
has been done in the area of development length (Tepfers, R. [98], Jirsa, J; Lutz,
L.; Gergely, P. [99], Orangun, C.; Jirsa, J.; Breen, J. [1 00]). The radial stress in
the concrete surrounding a bar being developed can be regarded as a water
pressure acting against a thick- walled cylinder having an inner diameter equal
to the bar diameter and a thickness "c" (the smaller of the clear bottom cover cb
~ ~J
or half the clear spacing C5 to the next adjacent bar). Based on a comprehensive
review of a broad range of test results, the following equation fort he development
length ( ld ) in terms of the stress in the bar at the critical section ( fs ), the bar
diameter (db}, concrete strength ( f c), cover (c) to diameter ratio, and transverse
reinforcement amount ( Atr ), yield strength ( fyt) and spacing ( s) were proposed
by Orangun, Jirsa and Breen [1 00].
=
A modified form of this equation in terms of a series of modifiers is the basis for
the recent changes in splice and development length design provisions in ACI
318-89 [1 01 ].
In CCT- and CIT- nodes the reinforcing bars are under lateral pressure
from the compressive struts (see Fig. 2.53). When lateral pressure is applied the
vertical component of the radial pressure tends to be balanced by the lateral
pressure. The bond strength increases approximately in proportion to the square
root of the lateral pressure. In addition, the distance between the bearing plate
and the reinforcement bar, e, has an important effect as shown in the study by
Lormanometee (1 02]. Different experimental studies were evaluated to develop
a formulation for a possible reduction of the development length for a reinforce)
ment bar under lateral pressure. Only tests in which failure occurred before the
bars yielded were included. The lateral pressure acts similar to the action of
transverse reinforcement. The overall strength of a splice with transverse
reinforcement and lateral pressure can be expressed as follows:
130
lateral pressure
=
=
(f'c) 05 (1.2 + 3 c I db + 50 db I Is )
(f'c) 05 [Atr fyt I (500 s db)]
(f'c) 05 [( fn )05 (200 - e2 ) 11 000]
{ 1.2 + 3
3.0
The comparison with test results from Lormanometee [102] and Schmidt-Thro,
Stockl and Kupfer [1 03] are shown in Fig. 2.55 and the statistical data is shown
in Table 2.18. The proposed relationship is conservative for all except one of the
test results and is generally quite conservative. A multiplying factor of 1.25 is
required to make the results consistent with the current ACI and AASHTO
expressions which indirectly introduce a <1> factor as 1 I <1> = 1 I 0.8 = 1.25.
L. '
131
Lateral presssure
Figure 2.54: Lateral pressure and the distance "e" to the reinforcing bar
Development length anchorage details include straight and hooked bars.
For these details the designer must check the development length requirements
of the tension tie reinforcement. For CCT-nodes confining reinforcement had
only a low effect (=2%). Similarly, for the CTT-node in which the transverse reinforcement anchorage hooks were turned nearly parallel to the longitudinal bars
(but not closed), the ultimate load decreased by only a maximum of 4% in comparison with closed confined reinforcement.
By using hooks instead of long bars for the anchorage, the ultimate load decreased by 8% for CCT (specimen C2 and 02) and for CTT (specimen HFSRA) nodes. Using a transverse U for the second tie in en-nodes provided lateral
confinement, but prying action at the 90 bend can produce splitting cracks. In
order to control splitting cracks of the end cover it is suggested that the
longitudinal reinforcement be extended a short distance (=s/2 or 2 in.) past the
transverse reinforcement.
Table 2.19:
x,: Column 1
Mean:
Std. Dev.:
Std. Error:
Variance:
Coef. Var.:
13.275
1.967
1.19
1.935
129.532
MaxilllJm:
.91
5.17
85.14
0
0
0
132
('I')
,.II
Q.)
.r:.
Q.)
sc:
.r:.
E
.:::
Q.)
a.
E
.:::
Q.)
a.
Q.)
Q.)
Q.)
c:
Q.)
0
0
0
tt
C\1
-.......ec.n
.......
...
Cl)
:J
en
en
Cl)
...c.
-...as
...,
Cl)
0
0
0
,.-
A.JoaLn
as
1 ~uaUJpadxa
Aq 4~5uaJ~S , U! 4~5uaJ~S
Figure 2.55: Comparison of a theoretical approach and test results for the
development length of straight bars with confinement from bearing
plates
L ,
133
2.7
Model optimization
Since strut- and- tie models are lower bound solutions for the actual load
carrying capacity of a structure, any correctly formulated and correctly detailed
strut- and- tie model should safely carry the design loads applied. In many cases
different models can be developed for the same external load configuration.
Doubts could arise as to whether the most efficient model has been chosen. In
selecting the model, it is helpful to realize that loads try to use the path with the
least forces and deformations. This simple criterion for optimizing a model may
be formulated as follows:
=
=
This equation is derived from the principle of minimum strain energy for linear
elastic behavior of the struts and ties after cracking. The contribution of the
concrete struts can generally be omitted because the strains of the struts are
usually much smaller than those ofthe steel ties. Since reinforcing ties are much
more deformable than concrete struts, Schlaich et al [28] propose that the model
with the least and shortest ties is the best.
As a more general approach for model development the following considerations are important constraints:
ease of fabrication
equilibrium
ductility
serviceability
.!
In many cases, practicality and ease of fabrication will have the greatest influence
upon the configuration of the design model. Models which resu It in detai Is that are
overly congested or difficult to fabricate should be avoided. The reinforcement
134
pattern for the D region should be compatible with the reinforcement scheme
used in adjacent portions of the structure. In order to satisfy the requirements of
the theory of plasticity, a model must be in equilibrium under the applied loads.
However, ifthe selected strut and- tie- model is to fully develop, the load carrying
capacity of the strut- and- tie- elements and the rotational capacity of the nodes
must not be exceeded before the ties yield. Furthermore, acceptable serviceability at usual working load levels requires that crack widths be limited by provision
of sufficient, closely spaced reinforcement in regions of high tension and hence
cracking. Attention must be paid to elastic analysis predictions of high tension
zones to ensure crack control reinforcement is appropriate. In addition to the
accepted standards for flexural reinforcement distribution and both minimum and
maximum bar spacings, minimum reinforcement to control shrinkage, creep and
thermal stresses should be provided.
r. 1, e,
I,
It
ld
e,
et
=
=
=
=
=
""
eN
=
""'
r. It et + r. f 1d eN
minimum
=
=
=
=
0.0
1.0
2.0
3.0 (reinforced TTT-node)
L_
r -,
L~
2.8
0.5 + 20 / (f'c)O.S
two
functions was chosen as shown in Fig. 2.56. The basic efficiency factor should
be taken as 0.8 for concrete compressive strengths up to and including 4000 psi.
For strengths above 4000 psi, the efficiency factor should be reduced continuously at a rate of 0.05 for each 2000 psi of strength in excess of 4000 psi, but
the efficiency factor should not be taken less than 0.65.
fee
ve f'c
ve
0.8 for
ve
ve
0.65 for
135
136
This basic efficiency factor can be used for checking compressive fields and short
struts within unconfined or !lightly confined nodes. as well as applications where
the compressive struts act over undisturbed or uncrushed concrete as occurs in
many wall type applications where no tensile cracking is expected.
ved
There are many reasons why the efficiency factor for compressive diagonals is
less than the global efficiency factor for unconfined nodes and undisturbed
compression fields. The web strength might depend somewhat on
the stirrup spacing in the longitudinal direction and the resultant control of inclined
web cracking. In addition the effective strength of the web may be reduced
because of cracks developed in early loading stages and having directions other
than that of the final cracks [21, 28, 62, 65, 66]. Finally, in beams and girders the
compression zones are highly concentrated and the struts in the web concrete
have a corresponding concentraion of load which may lead to more local failure
of the concrete at a stress level which as an average over the web is less then
the effective compression strength in more uniform compression fields.
137
(a)
The term "(1 - sld)2 " reflects the reduction in effectiveness of spiral associated
with increasing spacing of the spiral wires.
(b)
For nodes confined with orthogonal reinforcement such as closed square
hoops and with longitudinal reinforcement to anchor the corners of the hoops:
fe3 = ve f'c (A/Ab) 0 5 + 2.0 (ACJJre I Ab) flat (1 - sld)
:::;
2.5 f'c
(c)
For nodes confined with orthogonal reinforcement such as closed square
hoops but without longitudinal reinforcement anchoring the corners of the hoops:
fe3 =ve fc (A/Ab)0 5 + 1.0 (Acore I Ab) flat (1 - sld)
:::;
2.5 f'c
-I
' )
ve
f' c
A
=
=
=
~
Acore
As
flat
fy
d
s
=
=
=
=
=
=
fs
=
=
Jl
138
8
N
r---~---+----~--;----r--~~--~---+0
Cl)
Cl)
c::i
c::i
efficiency factor
139
2.9
Generally the development lengths for straight bars and for hooks should be
taken as recommended in ACI318-89 considering such effect as concrete cover, bar
spacing, and transverse confining reinforcement. Since the ACI 318-89 provisions
neglect the often beneficial effect of local bearing pressures such as occur at regions
where direct loads are applied or direct supports are provided, such local confinement can be considered if for design purposes the development length for straight
bars is computed as:
:_
3.0
0
6.0
The proposed formula take the lateral pressure into account whenever the
distance between the closest bar surface and bearing plate, e , is 14 in. or less ..
'
_j
140
i_
CHAPTER3. PROCEDURES
3.1
For many
not exceeded at any point before the assumed state of stress is reached in the
rest of the structure. This is especially important for the main members, which
carry a significant portion of the load. According to Schlaich et al. [28] it is
desirable that the struts and ties follow the elastic flow paths closely with a
deviation of at most 15 from the elastic principal stress directions. The proposed
design recommendations are applicable to either prestressed or conventionally
reinforced concrete members. The general assumptions for the application of the
strut- and- tie- model in the design procedure are:
Yielding of the reinforcement is required prior to concrete- or anchorage
failure
the ties transfer only unixial forces and neglect dowel action, aggregate
interlock, tensile strength across cracks etc.
141
__ j
142
~----~-------~)
l_
Figure 3.1:
,,
L.
143
3.2
I
I
I
I
I
I
I
I
1
I
_j
I
I
I
I
_j
Figure 3.2:
p
I
II
-1
'lj
~ ~v
9'
1\
I
I
'(
IJ
I'
0 -~~
144
In orderto be consistent with the factored load design methods of AASHTO and
ACI, load factors must be incorporated into the force calculations and ra factors
must be incorporated into the resistance calculations. For concrete compression
struts" factors as used in concrete column design seem most appropriate.
ra
f' c
b)
s;
ve
ve
0.8 for
ve
ve
0.65 for
s;
fZI
0.6 fee =
fZI
0.6 V ef'c
s;
fZI
f' cce
f'cce
)]
s; 2.5 f'c
longitudinal reinforcement
,__)
c)
(J
=2 fsAsl (d s)
145
!l
=
=
=
fs
~=
Acore =
A/Ab :::.;
1 S Acore/~ S 3
In determining the spread or diffusion of concentrated forces it is necessary to
assume or define a strut diffusion angle, the angle with respect to the strut axis
at which the com pression force spreads out from the edge of a bearing plate. For
the strut diffusion angle for heavily loaded members or under bearing plates
(anchorage zone) the following proposals are given. MacGregor [62] proposes
a diffusion angle of 15 degrees. An elastic finite element analysis by Burdet [42]
shows for various ratios of bearing plate width to compression field width from
0.1 to 0.9, diffusion angles vary between 27 to 22 degrees. The experimental
study by Sanders [43] gave somewhat lower values. Figure 3.3 show the various
approaches with the proposed equation (see Fig. 3.4)
diffusion angle [deg]
a
h
12 + 3/ (afh) 05
lead to conservative answers when compared with the elastic finite element
analysis for cases having ratios alh larger or equal to 0.15. Hno other information
about the a/h ratio is available, the diffusion angle can be estimated for design
by using an apparent diffusion angle of 21.8 {slope 2 : 5).
146
30
1!1.
25
D
C>
'C
......
.!
D
c
co
.._
20
----
.:.
15
"'\
0.1
0.2
0.3
a
0
!
~
it
10
-.
experiment
angle= 15
FEM
proposal
I
I
0.4
0.5
Figure 3.3:
24
.....
D
C>
'C
.!
C)
22
"
20
c.1l
18
-:e
16
:::J
14
:::J
--...
12+3/..J(ath)
..
c
0
0
---a--
~
..:..
"'W
ell
12
0.0
0.2
0.4
0.6
--.-
0.8
Figure 3.4:
1.0
L__,
,-
147
3.3
~J
l
c
'
(_ J
Figure 3.5:
The dimensioning is a check against the yield strength of the reinforcing bars and
prestressed tendons
~)
fy
'~
~fpy
=
=
As
AP
$
-1
=
=
148
Prestress forces are to be applied to the strut- and- tie model as external loads
(external force pair) with friction forces in the transmission zone (Fig. 2.31) in the
analysis and dimensioning. Only the available remainder of the yield strength
above the effective prestress force can be used for carrying tensile forces from
the strut and tie model (internal forces) (see Table 2.16).
After selecting the required spacing for the reinforcing bars and prestressing
tendons the width of the tie is determined as the outside dimension of the reinforcement layers. The width is necessary for dimensioning the node regions.
For those instances when it is desirable and permissible to count on the tensile
strength of the concrete to carry equilibrium forces where no progressive failure
,--1
seems possible, the following approach can be applied (width of the tension tie
assumed as 1 in.)
3 ..JfC h
depth of the tension tie
If the tensile stress field is crossed by a compression field, the reduced biaxial
L
3~
Ji>//1
0.25 f'c
Figure 3.6:
1.0 f'c
The maximum angle "()" between the compression and the tensile field is
arctan
~/ 24
For the resulting tensile- compression forces a parallel bounding failure curve
can be assumed.
;r
149
3.4
D-reg ions usually contain either smeared or singular nodes. The singular nodes are more critical and need more attention. The following dimensioning
and anchorage requirements must be applied to either smeared or singular
nodes. The stress peaks in smeared nodes are less critical because a greater
amount of surrounding concrete is normally available. For multiple, widespread
reinforcement layers it is difficult to choose the node width (Fig. 3.7).
Ca/2
~/2-:.---w2/2-..l
w1 = w2 coscl>l
Figure 3.7:
150
For the concrete efficiency factors and the anchorage requirements in smeared
nodes the rules for singular nodes (CCC, CIT, CCT, ITT) should be applied.
()
fee
ve
=
=
f' ce
ve f'c s 2.5 f'c
concrete efficiency factor
0
f' c
a)
b)
ve
0.8 for
ve
0.65 for
fc S 4000 psi
4000 < f'c < 10000 psi
Confined nodes
()
f' cce
f'cce
)]
2.5 f c
longitudinal reinforcement
=
flat
=2 f
7000 psi
151
c)
s;
G>
fbe
Ve f'c (A/Ab) 05
4
A!Ab s;
d)
fbe
(J
s;
2.5 f'C
stresses may be taken into account if the simultaneously acting transverse compressive stresses are considered reliable. This may be particularily appropriate
if supplementary transverse prestressing is applied.
When threedimensional compressive strength is appropriate
s;
G> fc3
fc3
s;
2.5 f'c
The dimensioning for CCC-nodes based on the proposed strut width ap(J
proaches given in Section 2.6.1 can be determined as shown in Fig. 3.8. For a
more general application with borders of the compression strut assumed to be not
parallel, the strut width "w3" at a certain distance "y" from the concrete surface
can be computed by using the proposed geometrical approaches shown in Fig.
3.9.
In a CCT-node where the tension reinforcement is welded or bolted to the
anchor plate, the stress configuration in the nodes is then similar to those in CCCnodes. A smooth surface of the tie where it crosses the node is better than high
bond from deformations because strain compatibility with the bonded bar will
tend to crack the node's concrete. The proposed CCC-node strength can be
applied forth is kind of loadtransfer.ln addition the anchor plate has to be checked
for bending strength and the welding connection with the tie must also be
checked.
effective width of the tensile tie is governed by the dimensions from the inside to
the outside reinforcement layer (wT). With a single layer of reinforcement wT is
taken as the bar diameter as shown in Figs. 3.1 0 and 3.11. With multiple layers
of reinforcement wT is taken as shown in Fig. 3.12.
:J
=~~
I=
I I I I Ic f I I f ll
152
a'
l.
i..
C1
C1
C1' I a'
=C1" I a"
C4 = C3 * cos 4l3
C4 = C2 * cos 4l2
a (tan 4l2 * tan 4l3)
80: - - - - - - -
Figure 3.8:
153
a
w4= al2
wc3 = w3 + 2 w3'
wc3 w3 + 2 w3"
w3'=
4 cos <l>3
_j
a = arctan [a I (2a')]
8 cos <l>3
)
Figure 3.9:
154
cr1
cr2
::;;
A/Ab
s;
fee
=
=
v e f'c
0.7
::;;
0 f ce (AlAb)05
0 f ce (A/Ab)o.s
::;;
2.5 f'c
::;;
2.5 f'c
The efficiency factor for the concrete compression strength should be taken as:
ve
0.8 for
Ve
ve
0.65 for
r--,
L_
[_)
I~- - --------~
~----,
wT= db
w2 = w1 sin <J!+ wT cos <P
Figure 3.10: CCT-node with single straight reinforcement bar
r~
155
The nodal zone must also fulfill the requirement for minimum development length, concrete cover and bar spacing limits. Test results [39] show that
vertically oriented hooks decrease the ultimate load of the CCT-nodes by 4 to 8%
as compared to straight bars with full development lengths. This decrease is
probably not significant given the other uncertainties in the deign process. The
advantage of hooks is that the required anchorage length can be minimized (Fig.
3.1 1}. With multiple layers of reinforcement, the available ld can be taken from
the intersection of each bar layer with the nodal zones (Fig. 3.12).
'
__ _i
l--/2--.. . .
--W1/2----..;
t--- - - - - a a
----~d~--too~l
wT= db
w2
156
l_
,1
l-\.
C1
w1/2
~I
w1/2
=
=
w2 =w1 sin 4>+ wT cos 4>
wT n db + (n-1) s
n = number of reinforcing bar layers
s clear bar spacing
\_
I_
157
cr1
:::;
A/Ab
:::;
fee
=
=
2.5 f'c
f'c
0.7
The efficiency factor for the concrete compression strength should be taken as:
ve
0.8 for
ve
ve
0.65 for
Test results [38] shown that the outside layers of reinforcement close to
a surface of the member are the most critical. Major cracks which initiate at the
surface and generally follow the theoretical strut angle decrease the bond
strength. Reinforcement should be provided across all planes of weakness to
control cracking. Confining reinforcement normal to planes of hooks and bends
is especially important [38]. Fig. 3.13 show singular tensile ties and Fig. 3.14
show multiple tensile ties for CIT- nodes .
. J
158
T1
I~-----------~~ --------~
.,..__jd
w1 = n'
w2 = d~
d~
or conservatively: w2
=0
;__
159
d'
~r-
T1
wc/2
)<C2
"' >
s'
wc/2
I~----------w1 --------~
w1 = n' d~
w2
=n"
d~'
160
L_:
L_
.
..
L__:
...
...
Figure 3.15: Special anchorage devices (from Ref. [1051)
r--
t"
161
__ )
..
_j
_j
._.J
+
Figure 3.16: TTT- node with looped bar
.,
_)
16.2
51
f'ct
R
L__
The equation are based on the assumption that the working stress in the curved
bar does not exceed 34 ksi.The stirrups can be omitted when the radius oft he bar
is large enough so that the cover concrete will be sufficient to supply the radial
tensile force.
R
In order to prevent splitting in the plane of the bars, Leonhardt [1 06] proposes
some minimum cover unless transverse reinforcement is provided (see Fig.
3.17).
[-
163
-stirrups
"
stirrupcurved tensile tie
Figure 3.17: Dimensions for curved tensile ties (from Ref. [1 05])
164
LJ
j:
L__ '
4.1
Detailing aids
Recent advances in the understanding of the behavior of concrete
o-s
= B = 4.2}
' J
Serviceability lim it states: These involve losses of the functional use of the
structure but not collapse of any part of the structure.
(1 00 year lifetime probability= 1o- 3
165
,_j
= B = 3)
166
Special limit states: This class involves major but repairable damage to the
structure due to abnormal conditions, such as long-term physical or
chemical instability
(1 00 year lifetime probability== 1o-3 = B = 3)
For normal concrete structures the determination of acceptable levels of
safety against occurence of each limit state are carried out by the building code
or design specification authorities. They specify the load combinations and
safety factors to be used in checking the limit states.
Since many repetitive computations are necessary to proportion concrete
structures, handbooks containing tables or graphs of the more common quantities are available from several sources. The American Concrete Institute and the
Prestressed Concrete Institute [1 07] publishes its Design Handbook in several
volumes, the German Concrete Group publishes its yearly "Betonkalender"
[1 08], the Concrete Reinforcing Steel Institute publishes the CAS I Handbook
[1 09]. In recent years, specialized computer programs have been replacing
design aids in many applications [11 0].
Detailing consists not only of the preparation of plans giving concrete
dimensions, reinforcement placing drawings and reinforcing bar details, but incorporates the whole thought process through which the designer enables each
part of the structure to perform safely Linder the various limit states. This chapter
gives some background for typical application of the strut- and-tie model. It is
intended to assist in establishing design parameters for some specific applications. These design guidelines should help a designer develop and dimension a
strut- and- tie- model and apply the model to different situations. This chapter
draws on the analytical and experimental results presented in the earlier
sections. It uses these results to develop design procedures for concrete
structures. Overall, it should lead to a better understanding of the force flow in
D-regions and the designer should have substantially improved knowledge
regarding the design process.
167
168
4.2
Types
The detailing process forD-regions begins with isolating the D-regions
from the B-regions in a structure and development of a preliminary strut- and- tie
model. In order to find the appropriate strut- and- tie- model the load path can be
traced, general knowledge of appropriate models can be used, or in complex
cases results of an elastic finite element analysis should be used.
All the design factors specified in this section are based on the 28-day
I__
L;
smooth wire, smooth bars and deformed bars. The yield strength varies between
120 and 270 ksi [55]. Am inim um amount of reinforcement is necessary to ensure
distributed cracking and should be placed to avoid infrequent wide cracks. In
critical cross sectional areas, crack prediction formulas can be used to distribute
reinforcement to avoid wide cracks.
L__
169
4.3
Figure 4.2 shows several different examples which will be developed in subsequent subsections. For clarity in presentation no nominal resistance reduction
factors (<P) or load factors have been used in the following examples. For actual
design purposes the nominal concrete strength
11
with the appropriate "<P" or resistance reduction factor. The design loads must be
increased with the appropriate load factors. Since these vary from code to code
and with actual applications, they are not included herein in the interest of
simplicity.
Nodes shown in Fig. 4.1 can occur in the different examples. Equilibrium
must be established in the nodes. The forces depend on the choice of their
position and are known from the boundary conditions of the B-regions. If nodes
with more then 3 forces occur the principle remains the same.
CCC
CCT
I C2
C1/
yn
'
C1 j!Zc3
en
TTT
T~
C1
T2
T2
T3
C2
T2
~1
AC2
c4-T
C3
Figure 4.1:
4Jc2
T1
Types of nodes
~1
T1v'"
T3
~1
T2v.
170
~-----------------T--------~~~-----1
I
I
I
I
J
t
I
Bregior>
P-<"llion
~T:-r
IP..t~~l
~~
from a column:
__j_
I
I
'~
D
=r:
I
I
I
:p
f... -1
L___
t.~
Anchorege zone:
I '-II
I
I
I
,_,_
D-norilioft , ,.s
I
I
I
l ',
-~. t,.5h~
!I
-l f-
',
I
I
I
I
........,.
~~ecntrk:tty
rl
i "~11111111111111-!llllllllllllll;.e, i
l
I
I
I
I
I
L-----------------~-----------------~
Figure 4.2:
s.,119 ..,
I
I
l
UhI
l
I
t- --l
I
I
--------+-----------------~
I
I
I
I
r 1- ~. ::
I :
1 l : : :~.:-1
o....;,
I
I
l
IF=F'F"
,.shco.a
-------~-----------------~
r + r
,!.,..,!,!
1
l
I
l
-f i
I
I
I
r--------
I
,;ih
---}
t:l-no!llcft
I
I
1""1.Sh-,
--,
:--"1
T.
['"'
I
I
Corbels projecting
I
I
I
~I
C. ,
171
For a heavy point load located near a support as shown in Fig. 4.3 and
Fig. 4.4, the proposed strut- and- tie- model is a logical approach to represent
the flow of forces indicated by the elastic analysis results shown in Fig. 4.3. If the
load is applied at a distance from the support smaller then the height of the
specimen {x ~h), then a compression field between the support and the load
. provides the primary internal force mechanism. This is also the reason why no
traditional shear reinforcement is needed in the area between the point of the load
application and the support. The strut- and- tie- model chosen reflects that the
primary compression strut between the external forces is a bottle shaped
compression field. The highly loaded bottle shaped compression field can be
represented with the local strut- and- tie- models shown in Fig. 4.4. The local ties
(T1) are dependent on the force diffusion angle "<j)1" and the compression force.
The T1-force in the tension ties can be provided for practical purposes by using
equilibrium to proportion orthogonal vertical and horizontal ties. The detailed
calculation in Example 4.1 shows that the tie forces in the bottle compression
field can be of large magnitude and have to be taken into account.
The bearing plate forces "F 11 were divided into two individual forces with
separate nodes. The magnitude of each force is determined from the overall
analysis according to the proportion flowing to the left support and that flowing to
the right support. In order to get uniform compression in the struts the bearing
plate width has to be also subdivided into two dependent widths matching the
compression forces F' and F". The example shows that the new strut- and- tieangle based on the widths a' and a" for this case does not have a significant
influence on the strut- and- tie- forces. The difference is only about 1%. It could
have a significant influence if the two compression struts "Ca" and "C4" had to
carry similar forces and the bearing plate is much larger {see Section 4.3.5
anchorage zone). The inclined compression struts outside of the support region
are assumed to be at an inclination of 45 as traditional in shear design {<j)7 = 45).
'
Thus the shear panel length becomes equal to the height Z in panels to the right
of section cd in Fig. 4.4.
172
Design a beam end for the member shown in Fig. 4.4 to transfer a vertical reaction load applied within a distance x ::; d to a supporting column. The load to
be transferred is 200 kips. Member dead load is neglected for clarity in this
example. Use f c = 5,000 psi and fY = 60,000 psi.
The computational steps are:
Estimate member sizes and dimensions
Divide member into B-and D-regions (see Fig. 4.4}
Develop a strut- and- tie- model (see Fig. 4.4}
Compute the external forces
Compute the strut- and- tie- angle
Compute the strut- and- tie- forces
Dimension reinforcement for ties
Determine anchorage requirements
Check concrete stresses
To prevent large crack widths under working loads some arbitrary reinforcement spacing limits are applied in the final design layout in Fig. 4.7.
173
'
_,__
. "'
-,. - ,--~//_J_
.
I
. ,,,_..///,+
- -
...
"
IJ
I
'r///?fj/
'
I
;(
;X ;1// t
I:J;X:;<'J(//It
IY:';(.j.>(.'f-1111
I f
i-
)Z )(';,( '/.
I I I
!0~999~~~:
I/,"/;~'1-.XXXXY-
t h"'-A~XXX'x_~'
I /,"'-/1-XX'x_~'---
l/,/,1-,>(X'x."-.'-'-_-
/,11>< X'-'-~
Z/;Xx,
j)!.------
lI ~~---I
I
Figure 4.3:
'
I I
l II f. J(';<)()( Xi-
'>-,
~
J.
I Iff-'/.';<)()( f..
J,
""' -1'
.......
..._
)(.
'
.r
...-
....
174
r==t
B-region
_!_
Iiii
D- region
F = F' + F"
. L
at
lr-...
Figure 4.4:
175
=30 in.
External forces:
A= F (t-x) /I
A= 186.7 kips
F'=A
B=Fx/1
B = 13.3 kips
P=B
Strut and tie angle:
tan la = z I x =0. 75 h I x
tan la = 0.75 * 40/20
tan la = 1.5; arctan 1.5 =.56.3 (old aa)
cpa = 56.3
alh = 0.25
81 = 12 + 31-J(a I h) = 18
A 1-.--2
....
x_J
176
Internal forces:
a"= F" *a IF
a" = 13.3 * 10 1200 = 0.66 in.
a' = a - a" = 9.34 in.
a'
-------lIIJioolj
a/2- a'l2
,-~-
4>7 is chosen to be 45
177
_j
~1j
__
= 223.2 kips
C1 =Ca I (2 cos 61) =117.3 kips
Ca =A/sin <j>a
(Negligible change)
__j
C3
F"
C3
cs
... 0~;;_;-
C4
T3 = C4sin 62.9
T5=T3
=F" =
13.3 kips
cs
C7 = T31 sin <!>7 = 18.8 kips
C6
- ........
d
T3
T4 = C5 = 115.6 kips
_j
V .
178
Grade 60
=1.20 in.2
,-,
I
Grade 60
For the vertical component T1v two No.4 U stirrups were used in the heavy
shear span.
Additional #4 U stirrups are placed under the load and just outside the
support (see Fig. 4.7).
L_
L_
T3 = TS = 13.3 kips
As3 =AsS= 13.3/60 = 0.22 in.2
Use #3 U stirrups
Since q,7 is assumed at 45, spacing of these stirrups can be z 30 in.
However, such wide spacing is unwise since major diagonal cracks could
form between such widely spaced stirrups. Stirrup spacing should be
restricted to z/2 or 30/2 = 15 in. #3 is the smallest practical size. Use #3 U
@ 15"(see Fig~ 4.7).
For the continuing horizontal tensile tie T4 try the same No.6 bars as for T2
('flexural reinforcement)
As4 = 115.6/60 = 1.93 in.2 < 5- #6 = 2.20 in.2
I'
179
Anchorage requirements:
The horizontal #4 bars and the vertical #4 stirrups provided to take the
T1 tie forces should be well anchored by hooking the horizontal bars around
the stirrups and hooking the vertical stirrups around the bottom and top bars.
No special check would be required for a member of this size as such
hooked stirrups could be easily develo19ed.
The main 'flexural reinforcement (Tie T2) anchorage at the support
needs careful examination. Using the provisions of ACI 318-89, for a clear
cover over all reinforcement of 1.5 inches and with #4 stirrups, the effective
cover below and outside the #6 bars is 2 inches. The clear spacing between
the 3- #6 bars in the bottom layer is 2.88 inches. Thus the cover is greater
than 2 db and the clear spacing is greater than 3 db satisfying ACI 318-89
Sec. 12.2.3.1 d. Hence, a multiplier of 1.0 is used with the basic
development length.
ld
= 12.73 inches.
However, since the side cover over the hook is less than 2 1/2 inches,
ACI 318-89 Sec. 12.5.4 would require that stirrup ties be spaced along the
entire ldh at a spacing of not more than 3 db, or 2.25 inches. This would
'
-,
make placement of concrete very difficult and so hooked bars are not very
desirable here.
In order to illustrate the effect of confinement due to the bearing plate
and stirrups provided, the more complex development length equation will
also be checked.
180
1.25 db { f8 I [ 4 (
f: )
05
" ]
-50 }
ld ---------------------~~--------~~--------------
{ 1.2 + 3 c I db
+(
rs (
200 - e 2 ) /1
ooo]}
e2
0.5 in.
ld
= 8 in.
for a = i o in.
The initial estimate of the bearing plate size was based on a length a
10
inches and a width equal to the beam width b = 1 2 inches. The bearing stress at
reaction A is thus f n
= A I ab
= 1 87k I (10) ( 1 2)
bearing capacity. From the clear cover selected ( 1 . 5 inches minimum), the #4 stirrups
chosen for the Tlv ties, and the spacing between layers selected (s = 1.5 inches), the
values of e1 and e2 can be calculated as 2.0 inches and 4.25 inches, respectively.
ld
[ 1.2
+ (
3 * 2.010.75)
+ ((
+ ( v1560
181
If the main flexural reinforcement runs to within 1.5 in. (for cover) of the
end of the beam as shown in Fig. 4. 7 and Fig. 4.5, the length of bar available for
meeting the requirements in node a is 15-1.5 + (0.75 I 2) {cot 56.8) = 13.75
in. for the lower layer and 15- 1.5 + ( 0.75 I 2 + 2.25) (cot 56.8) = 15.22 in.
for the upper layer. Thus the bars in the upper layer clearly exceed the 14.9 in.
required for ld according to ACI 318-89. The bars in the lower layer provide only
92% of the required ld according to the ACI 318-89 provisions* but do provide
188% of the required ld when the local confinement due to the bearing plate is
considered. Hence these straight bars can be considered effectively anchored
as detailed.
*(There is certainly
182
ve
Normally the compressive stress in nodes need only be checked where concentrated forces are applied to the surface of the structural member; e.g. below
bearing plates, anchor plates and over supports.
Node _g (Fig. 4.4): CCT- node
See Fig. 4.5 for node geometry.
w2
=
=
=
wa
a
a
L__
10 in.
b
=
12 in.
56.8
Ca
=
224.4 k
2 db+ s = 2 * 618 + 1.5 = 3.0 in.
a sin <Pa + w2 cos <Pa =10 sin 56.8 + 3 cos 56.8 = 10.0 in. (See
Fig. 4.5)
crca
crca
cra
cra
cra
Ca I (wa * b) ~ ve f'c
=
=
=
=
}=12'
a= 10"
...
183
T1v
.J
2#4
_j
wa
>
-.c
w2
--=--------~
ccl-...--hr-r-ri-T""T".....
t._~_J
wa = a
Figure 4.5:
~Bearing
sin~a
+ w2 cos9a
plate
T2
184
Node .!1 Fig. 4.4): CCC node
See Fig. 4.6 for node geometry
a"
=
=
=
0.66 in.
12 in.
<I>a
56.8
62.9
Ca
=
=
223.2 k
200 k
C3
122.4 k
C4
=
=
C5
115.6 k
cr,
Fl(a*b)~Vef'c
(J'cwa
a:
arctan [a I (2a')]
wa
9.3 in.
or
wa
9.3 in.
a
a'
<1>4
10 in.
9.34 in.
/I
//~al
14.9 k
r-,
OK
28.16
OK
_j
'
185
;_ _j
L--~J
ct:
=
=
arctan [a I (2a")]
w4
crew4
crc3
C3 I (w3 * b) ~ Ve f' c
w3
=
=
crcw3
Sin
a/2
=
122.4 1 (5 * 12) = 2.04 ksi
crews
C5 I (w5 * b) ~ V e f'c
w5
a/2 = 5 in.
crews
115.6 I (5 * 12)
crew4
82.480
4.75 in.
OK
OK
OK
The assumption that w5 = a/2 assumes a hydrostatic type node. Note that the
computed uniform stress in C5 is only 1.93 ksi or only 45% of the 0.85 f'c assumed at failure in compression zones under ACI or AASHTO [3,4]. This indicates that the initial assumption z
=30 in ..
= 33.4 in.
Calculations
could be revised on this basis but since everything has checked OK at the lesser
assumed z , forces would be decreased. Therefore, this analysis is conservative.
Concrete stresses in compression diagonal struts:
Since this application of load resu Its in a heavy diagonal strut rather than a
compression field, it might be useful to check the main strut system for the criteria cr ~ 0.6 ve f'c = (0.6) (0.775) (5)
= 2.33 ksi
Since the nodes were checked for the concentrated Ca force, it is not
likely that C1 or C2 will govern. However, to check the level of stress in the
186
=wa + 2
a tan
,.,
187
a'
a/4
wS::al2
_l
l..,._aJ2+a"/2--l
w3=w5 =a/2
ct
=arctan a I (2a')
0: = arctan a I (2a")
' J
or
wa4 = a"cos 4 + a/2 sin 4
Figure 4.6:
188
I..- A
"(--
40
I
I.._ A
I
f-oolt----4o"- ----....,._---- ----0- region= 80"
Section:
A~
---------~
__2.375
---#4
8.25"
t
3.5"
~
8.25"
40"
8.25"
+-
1.5"
8.25"
,---
i
2.25"
t2.3a
'
~.25
Figure 4. 7:
Reinforcement layout
~--~
189
1. crack
2.crack
190
In tests, corbels display several typical modes of failure, the most common of
which are:
yielding of the tension tie
failure of the end anchorages of the tension tie, eitherunderthe load point
or in the column
failure of the compression strut by crushing or shearing
local zone failure under the bearing plate
When using a reinforcement tie hooked downward, as shown in Fig.
4.9(a), the concrete outside oft he hook may split off. In orderto avoid this problem
closed ties extending past the loading plate may be used, or straight bars may be
used but should be anchored attheirendsby welding them toacrossbarorplate.
lfthe corbel is too shallow at the outside end, there is a danger that cracking may
extend through the corbel as shown in Fig. 4.9 (b). For this reason ACI [4] requires
that the depth of the corbel be at least 0.5 d at the outside edge of the bearing
plate.
L ;
A total of eight corbels, divided into four series with concrete strength
ranging from about 6000 psi to 12000 psi, were studied by Yong, Closkey and
Nawy [114]. Only verticalldad was applied. The specimens which had no steel
reinforcement (fc
=10260 psi and 12630 psi) had failure at the interface of the
corbel and the column as shown in Fig. 4.8. The failure plane started at the
reentrant corner of the horizontal surface of the corbel and the vertical face of the
column. All the other specimens had almost identical behavior when subjected
to failure under a vertical load. A flexural crack (crack 1) first started at the
reentrant corner and propagated slightly into the column. Close to failure another
crack (crack 2) appeared at the inner edge of the bearing plate and propagated
at a faster rate than the initial crack towards the interface of the column and the
sloping face of the corbel.
L...
191
Figure 4.9:
192
ma,nitvde
1.ompressie trajKiories
or the
193
In the traditional approach to the problem, one would have relied on the
consideration of shear stresses. Indeed, corbels have often been reinforced with
diagonals, as shown in Fig. 4.11 , to take a substantial part of the shearing force.
The investigations of Franz and Niedenhoff [115] have conclusively proved the
inefficiency of this approach.
r---;-----J..--
I
---t--1
vi
(b)
(b) dlllgoi'IIIIIIJ.)Iitling
ld
194
The strut- and-tie- model fora corbel shown in Fig. 4.13 is proposed and assumes
that the concrete stress at the lower reentrant corner reaches the effective
concrete strength [28]. A typical CCC -node is found in the lower corner and the
tensile force in the upper chord can be computed without knowing the angle of
I__ _
inclination of the main compression strut. The flow of forces and typical CCCnode dimensions are shown in Fig. 4.14.
l_-
CCT-node
Tie
Strut -
L_
L_,
Figure 4. 13:
Figure 4.15 shows eight test results from [117] with a concrete range from 5690
to 12630 psi, six test results from [118] with concrete strengthts from 4200 to
5057 psi and one test from [113] with 5858 pst compared with the proposed strutand- tie- model. The statistical analysis forthe comparison is shown in Table 4.1.
The proposed model is generally conservative and reasonably accurate.
L,
195
'I
'c_,
Fv
Fh ....
=rd
, __!
: _j
CCC - node
C3
w2
C2
c_j
=___F_v__
b {f'C
196
2.0~-----------------------------------------------~
u == unreinforced
---t:J---
experimentttheory=1.0
>'-
experiment/theory
1.54-------------------------~----------------;
C3J
~=
-g-
ocQ1
-Q)
ro e
'Q1
::::> c.
E
.:::
><
Q1
6000
8000
10000
12000
14000.
Figure 4. 15 :Strut- and- tie- model results compared with test results
IL082
Std. Dev.:
X1: Column 1
Variance:
Std. Error:
1.155
1.04
Minimum:
Maximum:
Mean:
.93
Table 4.1:
1.442
1.024
Coef. Var.:
Count:
114.311
115
, 6.224
197
Example 4.2 :
= 10.5 in.
Since tie reinforcement often has to be placed in multiple layers, cover and
spacing requirements suggest that a conservative estimate be used for the
effective depth, d. The basic strut- and- tie model shown in Fig. 4.16 is
chosen for computing the strut and tie forces. The compression struts are
again assumed as "bottle" struts. The location of node b is uncertain and
_j
depends somewhat on the location of C4. After C4 is located the node b will
be chosen on a 45 angle inward from the reentrant corner.
198
l .. '
l_
C3
L _:
199
. . . ---h----....,,
1.5 h
-l
Fv
1.5 h cosB
_l
Figure 4.16: Strut- and- tie model for corbel projecting from a column
200
Internal forces:
C4 = Fv = 200 kips
=3.0 in.
J!umn bars indicates C4 should be at least 2 in. from the column face. Use
j_
4.0 in ..
'th the location of C4 at w412 or 2.0 in. inside the column face, the location of
de b will be chosen 2.0 in. inward from and 2.0 in. above the reentrant corner.
:s establishes the angle of inclination of the strut Ca:
Tan (jla
$a
Ca
=
=
48.6
1.133
Fv I sin $a
266.6
=
=
=
ali
8
T1
T1
=
=
=
=
a cos $a+ Fh
6.3 k
'cos (jla
= 0.55
+ 3 I .V 0.55
16.0
176.3 k
201
T1
=38.2 kips
T1 h = T1 sin <j)a
=28.7
kips
= 1.20 in. 2
= 3.60 in. 2
a= 5 in.
I_)
OK
wa
O'ca
= 5.40
OK
202
w4
= 4.0
w3
w1
O'ca
Ca I w1 b ~ 0.7 f'c
5.6 ksi
0'3
0'4
C3/w3b = 176.31(3.52}(12}
C4/w4 b
= 4.17ksi
OK
< 5.6ksi
OK
OK
L_.,
,,
I
l )
'-
'
203
Anchorage requirements:
l
L_j
'
Three closed #7 ties were used for the main horizontal reinforcement. Because
the tension tie in the strut~ and~ tie~ model is assumed to be stressed to the yield
strength intension between the loading plateandthecolumn, it must be anchored
in the node zone a and outside the bearing plate for that tension. The closed ends
of the ties should be sufficient positive anchorage. If straight bars were used they
could be welded to an angle or bar at right angles to the tie (see Fig. 4.17), or be
welded to a transverse reinforcing bar of the same diameter as the tie.
bearing plate
~777727ZZZZZZZZZZZA
~----~1.-------/
Figure 4. 17: Anchorage detail for corbel design (from Ref. [62])
...,._--12" _ ___,...,~I
204
6"
2 - #4 bent bars
15"
_l
Section: A A
closed ties #4
l__
t1
3@2.
7"
_ll
, s
27"-------.....,
205
A deep beam is a beam in which a significant amount of the load is carried to the
supports by a compression thrust joining the load and the reaction rather than
through flexural action. This occurs when a concentrated load acts closer than
about 2 d to the support, or for uniformly loaded beams when the span- to- depth
ratio, ''I/ d", is less than about 4 to 5. The ACI Code [4] specifies that deep beam
action must be considered when designing for flexure if "I/ d" is less than 5/2 for
continuous spans or 5/4 for simple span (see also Ref. [1 05]).
Cook and Mitchell [113] did some experimental verification and non liner finite
element analysis of a uniform loaded T-beam with a hole in the web. The finite
element analysis by Schlaich et al [28] and the experimental study by Cook and
Mitchell [113] show that tensile forces are acting especially around the corner.
This leads to the assumption that
discontinuity zone in the tension region. The stress concentration factor for an
infinite plate with a rectangular hole and subjected to biaxial stress is the highest
such factor for all different forms of openings [119]. For a finite-width member
with infinite thickness and
12 I (7- 5 v)
0.16 for concrete
1.93
For the corresponding case of a semi-infinite body, Tsuchida and Nakahara [121]
developed stress concentration factors. The values with Poisson's ratio of 0.25
and r
are:
r/m
0.5:
2.32
rim
0.8
3.3
For rectangular openings the following mathematical results, with specific data
obtained by computer, have been published:
' .J
<J1
=0 2
206
tension:
cr,
I
~r
tension:
cr2
b=a;
r/b
0.1
b=a;
r/b
0.3
4.88
2.76
207
Schlaich et al [28] proposed that, for this kind of problem, two separate
strut and- tie- models should be developed, each with a carrying capacity of 50%.
One model should follow the elastic principal stress trajectories with a diagonal
tension tie and the second model should have strut and ties parallel to the
borders. From a practical standpoint it is very inconvenient to place diagonal
reinforcement in many structures.
Test results from Shah [122] and Gaynor [123] for tests on reinforced
concrete in-filled shear walls with openings gives some indication that first
cracking appears near the openings. In order to prevent large cracking and for
crack control under service load it seems reasonable to round off the corners
(see Fig. 4.19: r/b ::::< 0.3) and for geometrical discontinuities subjected to biaxial
tension a quantity of diagonal reinforcement equal to about 1/8 of the orthogonal
reinforcement should be provided as an addition. In the general literature on
design [62, 105,113, 124] no detailed information about the required amount of
diagonal reinforcement is given. As shown in Fig. 4.20, such diagonal reinforcement follows the principal tensile stress directions closely and should be very
effective in controlling the reentrant corner crack width.
Figure 4.20: Principal tension trajectories and reinforcement for corner in tension
208
Example 4.3:
Design a deep beam with an overall depth of 16'8" and an overall length of
27'6" to transfer a vertical load of 500 kips applied 14'7" from the left edge.
The wall is supported on simple supports located 25'0" on centers (see Fig.
4.21 (a)). The supports are 20" X 15" columns and the wall thickness is 15".
There is a large hole 65 in. square located in the lower left corner. The hole
has 30 in. of concrete below it and its left edge is 30 in. from the center of the
left support. Concrete strength is 7000 psi and Grade 60 reinforcement is
used.
m
a
b
=
=
=
=
=
=
=
=
=
500 kips
300 in.
30 in.
65 in.
200 in.
30 in.
65 in.
160 in.
20 in.
15 in.
External forces
B
A
267 kips
233 kips
209
In developing a strut- and- tie model for this type of unusual structure,
it is very useful to consider the elastic stress pattern indicated by a finite
element analysis. A general pattern for a similar type of problem with the
load located considerably more to the right has been provided by Schlaich
et al [28] and is shown in Fig. 4.21 (b). Based on the pattern of elastic
stresses shown, it can be seen that significant tension acts in the diagonal
direction at the upper right corner of the opening and lesser tension acts on
'_j
the diagonal at the lower left corner of the opening. The thrust to the left of
the opening is skewed slightly to the right, inclining towards the opening's
upper left corner. Schlaich et al [28] have suggested two possible strut- andtie models for the left side of the structure as shown in Fig. 4.21 (c) and (d).
They suggest the left reaction be considered as split on a 50-50 basis
between these patterns. One minor problem with these suggested models
is that there is no tie required beneath the opening. Inclining the thrust
towards the upper corner of the opening in the section to the left of the
opening would result in a tie requirement beneath the opening. The part of
the wall to the right of the applied load is basically a straightforward case
with the thrust being transferred by a bottle-shaped strut to the reaction at B
and the lateral component of the strut force tied back by a lower tie. For this
particular example, variations of the models suggested by Schlaich et al
were adopted. For the portion of the wall to the left of the centerline of the
applied load, it i~ assumed that the load carrying capacity will be shared
equally (50-50) by the Strut- and- Tie Models shown in Fig. 4.21 (e) and Fig.
4.21 (f). The geometry of the models and the resulting strut compressions
and tie tensions are shown on each figure. Note that in Fig. 4.21 {f) C8 was
assumed as a compression strut but in the solution (performed using a
microcomputer program for a 20 truss based on SAP) was found to have a
very low level of tension. Similarly Til was assumed as a tension tie but
analysis indicates a small amount of compression. While the two models to
the left could be combined, it is easier to proportion reinforcement using the
two separate models. The much simpler section to the right of the load is
shown in Fig. 4.21 (g) with the right reaction, 267 k, and an equal part of the
load applied to a bottle strut and major tension tie.
210
718 h
Tan ~b
175/135
~b
=
=
=
=
=
52.35
a/h
91
91
175 in.
= 1.296
10/200 = 0.025
12 + 3/..J(a/h) ~ 25
12 + 3/..J(0.05}
=25.4 ~ 25
Use 25
Internal forces
Cb
8/sin cj>b
C14
T14
=
=
T13
Cb cos cj>b
=
=
(Cb/2) tan 91
= 337.2
kips
186 kips
78.6 kips
206 kips
L_
'-
211
_j
a= 20"
l_j
15"
m = 160"
... l5ook
15.
140"
...L
_I_
~~
l
:::;
___I
....-..1..
I
::::::
,.
:::::
~t:
' _j
I ___;
It
::::
..,1
b = 20.
a= 20"
15.
...
~ B
I~
I= 300"
-,
Figure 4.21: Strut- and- tie models for deep beam with a hole
c.J 111111111111111. . . . . ._ _ _ _ _ _ _ _ _ _ _ _ _ _ __
212
r"
Principal Compression
- - - Principal Tension
Figure 4.21: Strut- and- tie models for deep beam with a hole
213
A1
=O.SA
Figure 4.21 : Strut- and- tie models for deep beam with a hole
214
A2 =0.5A
___ - - - -
----ol
.
I I c =o.sc
2
I
I
i
:i
I
I
I
i
I
: iI
~-Tg
: i
I
A2 =0.5A
Figure 4.21 : Strut- and- tie models for deep beam with a hole
;__j
215
C1 = 104.1k
C2 = 110.5k
C3 = 120.8k
T1 = 134.9k
T2 = 25.6k
0"
C\1
:0
'
ro
("-...
'r"
'
I
I
65"
T2
155"
'
IT= 103k
ro
T-
216
a= 59.4
B = 13.8
Compression
C4 = 116.7k
cs = 164.sk
C6 = 123.7k
C7 = 111.3k
ca = -a.3k(T)
C9 = 154.4k
c1o = 91.sk
C11 = 145.7k
c12 = 103.ok
C13 = 145.7k
T6
~C4
... TS
.. ...-
T4
'
L---
L_,
/C13
45
-)_...
/c11 :c12
Tension
T3 = 103.ok ,--,
T4 = 103.ok
Ts = 103.ok L ,
T6 = 206.Qk r
T7 = 203.7k
T8 = 206.9k '
T9 = 203.8k ,T1 0::;; 116.3k
~ T11 = -9.0k(C)l_
.,.... T12 = 87.5k r-
T3
T=
:... 1Q3k
:
'
L:
L_~
116.3k
155"
L_
L_
217
IF = P + P = sook
t F" = 267k
'
-~
NodeC
Cb = 337.:
C14=186k
\~,
\ \'C14
\
'\
k
.J
''
\cb'
'
V'\
\
.\
\
C)
'
C15 = 168.1
T13 = 206k
T14 = 78.61
\\~
'\.C15
\ \
\~\T14
\
\
''
___ j
T13
\
\
\
,.
q>b ''.\
Node b
h =200"
\
B = 267k
135"
....a
03
--
C1
---I
----C7
CST 1
-----
I
I
~~
1
I
\..
T7
I .--------.
\ \ ',C14
\
''
\.
\
\
'
'\
\C14 T14
\
1c3
',C15
'
\
'
'
'
'
T4
',
\, T14
\,
T2+T3
\C14
'\,
' , f"'14 \
~
T13
T2
'
.\
/
/C11
I
.
'C15
I...._C2 + C12 '
I
I
I I
I I
I I
I I
II
I
I
TS
c41
~'
\'
T11C /
C6/'
"
',
\
'
\
\
\, \
',,\,.
'
A= 233k
\
B = 267k
219
206 kips
As = 206/60 = 3.38 in.2
Use 6- #7 = 3.60 in.2 in lower right section (See Fig. 4.22)
T13
T14
T14v
T14H
= 157.2 k
= T14 cos cj>b = 157.2 cos 52.4 = 96 k
Asv = 96/60 = 1.60 in.2
=
T14 sin cj>b = 157.2 sin 52.4 = 125 k
78.6 kips X 2
'
.....
::J
....
.....
f)
....
,....
:\1
'c;t
::::J
g>
~1
As
0.29 in.2fft.
.I-
(See Fig.
220
Anchorage requirements:
#7 bars for T13 ties at node b (Fig. 4.22) (CCT node)
The 6 - #7 bars required can be efficiently placed in 2 layers of 3
bars each with a clear cover of 2 in., a clear spacing greater than 2 in.
and the vertical #4 bars bent below the #7 bars. This results in 2-112 in.
of concrete below the bars. From Fig . 3.12 and the dimensions shown
in Fig. 4.21 (a), assuming 2 in. clear cover over the tails of the #7 bars,
w1 = 20 in. and the length available to satisfy ld requirements is
w1 +5-2 = 23 in. From ACI 318-89 Ch. 12, ldb = 0.04 Ab fy ..Jf'c =
(0. 04) (0.60) (60,000) 1 ..J7ooo = 17.2 in. In a 15 in. thick wall with
clear cover of 2 in., two #4 bar vertical curtains, and three #7 bars in a
layer, the cover is 2-1/2 in. and the clear spacing is 3.7 in. which is
greater than 3d b. Therefore, a factor of 1.0 is used.
ld
OK
No hooks required. Anchorage foF the other end of these bars at nodes
(a) and (c) (Fig. 4.22) will be checked as part of the left portion.
#4 bars for T14 ties
ldb = (0.04) (0.20)
The required ld for these bars is short.
(60000) I ..J7000 = 5.7 in. Since cover is 4 db and spacing is 32 db
in one direction and 20 db in the other, ld = 1.0 ldb = 5. 7 in. > (.03)
(0.5) (60000) I ..J7000 = 10.8 in. which governs here. Clearly there is
no problem along the top or right edge of the wall. As a good detailing
practice to provide confinement for the main tension tie, the vertical #4
bars should be U type hairpins and enclose the #7 bars in the main
tension tie (see Fig. 4.22).
221
w2
=
=
=
=
=
crcb
crcb
cr1
OK
OK
From Figs. 3.9 and 4.21 (g), $3 = 37.6, $3' = 62.6, $3" = 12.6,
a= 20 in., a'= (2671250) (1 0} = 10.68 in.
Checking for w4 = a/2
= 10 in.
C4 =
Cb *cos $3
= 337.2
cr4
c4 I {w4 * b)
~ Ve
Check strut cb
cr3
=
=
f'c
=5.08 ksi
OK
at y =a= 20 in.
222
223
Use 4- #7.
Anchorage Beguirements
When checking the right side of the wall it was determined that for #7 bars, ld
= 17.2 in. At support A at least 23 in. > ld is available. The smeared nodes
at the right ends of the T5 and T7 ties require only normal ld embedment
past the node, as does the upper ends of the T12 and TS ties. In the more
critical cases at the left end of the T1 0 and TS ties and at the bottom of the T4
tie, positive anchorages are provided by looping the ends of the bars. The
orthogonal curtains of #4 bars provided for the T14 ties require no further
check as they have ample length to satisfy ld requirements.
L ....J
224
=75
err
Node. Detailing
Check Node (d) [Intersection of T1 T2, C2, T13]
continuous bars has satisfied T requirements. C2 has a force of 11 Ok and
must be basically equilibrated by the bends of the 4 #7 bars in the T1 tie.
For this 55 bend, a standard inside bar diameter of 6 db would result in a
I
bend contact area of approximately (7) (0.88) 1t/4 = 4.84 in. for each layer. An
extremely conservative estimate of node pressures would be:
crcd = c2/ w1 b = {11 0) I (4.84}(15)
~~
I..
~~
. I.
~s
.. I..
-~
..
-...
a___.,.
\l\
2
\. '\.\.
.\
-r=;-
2 x 1I - #7 IBent
v
+-----1---1
)-2 X 2 - #7
'
\., \.
Looped
~--*-~
A'
"'-'--------1/1
1"'- /
r>,
-J-T~~-4~~~~
~ ~ 2,_B"J-t-1--t--f--+--+----J---J
2x1-k4
~
111111
..
Looped at
bottom
/ /
....,._2x3-#7
::
.. I
............,
mnn
2@
2x 2#7
w/#3Column
Ties@ 14" c/c
Bars shown as
2 x _ indicate
one set in each
face curtain
...
..-f ~
\.\
ooped
'
"-2 X 3- #5
62.5"
~
7
1
r-
2x2-#7
Bent up
N
N
(J1
226
L_:
,----!
fidelity
to
the
elastic
principal
stress
l_
227
Figure 4.23: Reinforcement layout for strut- and- tie- model ST1
(from Ref. [7])
,.,,..;
I
,.. ,,..
t SI'ACE'
Figure 4.24: Reinforcement layout for PCI detail (from Ref. [7])
failure
228
!._
229
230
::;:;:::;::
::::
;-::
;:;::;:
;:;c:
:;:;:-:-:::
~-
:;:;:;
"'
.,.,.
z.,:,:,
r--
i1
JtA ' 1
::::~
llfji''
!{: 1{::
It:
~L~
::;:;:::~
l'l.: l:l
-~-!
:I'!?
II~~:::;:;:;:;::;:::::::::::
1::;:;:
!:~ljlr
:;:;~
:~}:
!;T3:::
'IS
r:::)::::::~:
:-::::
;:;:;::
:r2 = T4 =A
T
i
l
T1
=A I sin cp1
231
T
i
__j
'
T1
.4~A q,1
l
Strut and tie- forces:
Ca =A I sin
C1
9a
T1 =A/tan <:>a
C2 =T2/ sin 2
C3
C4 :T41sln~
:-
T4
= C1
TS:TS + C7 cos 7
Control : CS
= TS
Figure 4.29: Proposed strut- and- tie- model for dapped end beam
232
L_
$a
fl
= 55"
Proposed geometry:
lla<:!1.15
Figure 4.30: Proposed strut- and- tie- angle for dapped end beam
233
2.0,~--------------------------------------------~
c = concrete crushing
-c- experiment/theory
-e- experiment/theory = 1.0
t = test was stopped
1
~1.5~==~---===~~~--------------------4
:+::::;Q
ro m
'-..c:
"'ro~ ~ 1.0~------------~------~----------------------r
-E
ca._
E 0>
-c..
-x
Q)
0.5+-------------~--------------------------~
specimen
Figure 4.31:
Table 4.2:
Column 1
Mean:
Std. Dev.:
Std. Error:
Variance:
Coef. Var.:
Count:
1.37
.13
.37
---
0.09
Minimum:
Maximum:
Range:
Sum:
Sum Squared:
#Missing:
1.21
1.52
.31
5.47
7.55
234
Example 4.4:
Design the end region of the dapped end beam shown in Figure 4.32 to carry a
concentrated midspan load of 300 kips. Dead load of the span may be neglected in
this example. Dap details are shown in Figure 4.32. Use
f:
300 kips
200 in.
30 in.
5 in.
14 in.
16 in.
10 in.
15 in.
5 in.
~~-
~-.
The beam may be divided into D regions near the ends and a central B region (a D
region may be used under the load but is not checked here). The B regions will be
L_
External forces:
150 kips
r-
3/4 h
22.5 in.
r-
55
<P2
</)4
<P2
</)3
=
=
=
=
235
45
45. From the geometry of Figure 4.30, one can compute
= 52.1'
11.75 - 10.72
11.75
= 5.0
Internal forces:
Ca
150 I sin 55
C1
C1
148.5 kips
183.1 kips
To find <P2
,,A
,I
;
;
,'
,l' /
I
45? , ,
/
I
~1.677~552.39
2.39
~''I ,'
I
8.33 Tan 45
=8.33
a/6 =1.67
cir.28.33
a13+ e= 8.33
To find </)3
16"
236
A~region I
B- region
~1.5 h--1
~1.5 h---'. 1
1 - - - - - - - - - 1 =200 .. - - - - - - - - - - ; -
0- region
.,..
16"
,.. ..cs
.CV
~~)
c2
T3
T~/
14"
j_
,../
T7
/
T2
---cs---,..
,..
TS
.., ,..
/
~
/
~
c7
T9
/
/
j/
T&
l~
//
TB
..
{b)
. 5"+6"
Figure 4.32: Strut- and- tie- model for example: dapped end beam
237
,_,J
_j
_ _)
T1
T1
=
=
A I tan q,a
105 kips
T2
T2
=
=
C2
C2
=
=
T2 I sin $2
57.0 kips
C3
C3
=
=
T3
T3
=
=
T1 +C2 cos $2
140.0 kips
T4
T4
=
=
<f>6
C4
C4
=
=
T4 I sin $4
194.8 kips
30 - 3 - 3.5 = 23.51
C5
C5
=
=
C3 cos $3
140.0 kips
T5
T5
=
=
T4 I tan $4
137.7 kips
T7
T7
=
=
C3 sin $3
12.2 kips
---------0
/
<1>6/
//"'----
T7
11.75"
90- <I>~<
( J
//<t>6
\.__/
_)
'i'
'l
''l
'
,-,
__ )
$6
C6
C6
=
=
T6
T6
=
=
T7 + C4 sin $4 - C6 sin $6
151.1 kips
26.6
________ )_
238
C7
C7
=
=
T8
T8
T5 + C7 cos <1>7
288.7 kips
C8
Check C8
=T8
by method of sections
=289.5 kips
Check OK
T9
= 149.9 kips
= A = 150
Check OK
105 kips
1.75 in 2
6. #5
1.86 in. 2
T2
As
=
=
45 kips
0.75 in 2
4 -#4
0.80 in. 2
T3
As
140.0 kips
2.33 in2
:::;
8- #5
2.48 in. 2
T4
As
=
=
137.7 kips
2.30 in 2
:::;
12- #4
TS
As
=
=
137.7 kips
2.30 in 2
4-#7
2.40 in. 2
2.40 in. 2
lc
239
T6
As
=
=
151.1 kips
2.52 in 2
:s;
14- #4
T7
As
12.2 kips
0.20 in 2
::::;
1 - #4
288.7 kips
4.81 in 2
:s;
8- #5 + 4- #7
150 kips
2.50 in 2
:s;
14- #4
T8
As
T9
As
2.80 in. 2
0.20 in. 2
4.88 in. 2
2.80 in. 2
A possible bar arrangement is presented in Fig. 4.33. Note that the T6 bars
are run full height and provide substantial excess for the T7 bars. Note also
that the #4 bars required for T6 and T9 are provided as Groups of ciosed 3 W stirrups and 1 - U stirrup. The U stirrups provide a desirable transverse tie
completely across the bottom flange. Where W stirrups are provided a short
U is desirable on the bottom flange. (A W stirrup is a four-legged stirrup
CITI)
240
Anchorage requirements:
The very congested conditions in the dap make it difficult to effectively
anchor all the reinforcement. The 8 #5 bars provided for the T1 and T3 reinforcement should have positive anchorage by closed loops at the support. In addition,
the lower layer should be welded to the bearing plate, if possible. The minimum
development length for a Grade 60 #5 bar under ACI 318-89 provisions would be
14.5 in. There simply is not room available in the small dap to rely on development length, unless one relies on the confinement present, since the T1 force
must be fuHy developed above the bearing plate which is only 10 in. long.
Similarly, the #7 bars provided for the T5 force need to be developed
within the
node at the bottom corner of the full depth section. This is also a
very congested corner. Under ACt 318-89, a standard #7 hook would have a
basic development length of 13.5 in. However, the 2" clear cover over the #4
stirrups provides 2-1/2 in. side cover so that a multiplier of 0.7 may be used
reducing the length to 9.5 in. However, the highly grouped T4 reinforcement
greatly reduces the width w1 of the node (see Fig. 3.14). It is highly desirable to
space stirrups at no more than 3db = 2.64 in. throughout the hook development
length to allow use of an additional 0.8 factor reducing ldh to 7.6 in. This can be
easily done by using the 3 stirrups required for the T4 tie at 1-1/2 in. on centers
and then introducing 2 extra confining stirrups at 2-1/2-in. spacing. One of these
can be counted towards the T6 tie force so only one additional stirrup is required.
en
CCT- node
241
w1
wT
wa
=
=
=
6.7in.
2/3 (1 0)
=
2.50 in.
2(0.625) + (2 - 1) (1.25) =
6.92 in.
w1 sin 55 + wT cos 55 =
(wa is the same as w2 in Fig. 3. 12}
O'ca
Ca I (wa * b) ~ v e f'c
O'ca
'
C.,)
_ _j
w1 (T4)
w2 (T5)
we
=
=
CTT- node
~0.75
* 6. = 4.5 ksi
OK
O'c4
O'c4
4.5 ksi
OK
242
2 #4 bars
3-#4 bar
8 #5 bars w/ looped
ends at support -
#4 stirrups
@3"
* f f f l: t t t f f f 9 9 = : f
welded to bearing plate
3 grouped #4W
stirrups 1t on
centers
1o
{T4)
Section:
4 #7 bars
vertical hooks
8 #5 bars
straight bars
,.. A-
A-A
3"
"
i3l.
4
30"
r--
_L
#5 bar
2"
T
#Sbar
#5 bar
#7 bar
1-'----
15" --~
243
';t'
'
1/2 of the diagonal or diameter of the plate minus the radius of the
'
'
~%
wedge plate
"{
~'
~~
-i!
~
0
'
-=
~
-!}
:i
~'
fb
A
F
0.85 F I A
244
If the plate cannot be considered rigid, it may be used but the effective bearing
area shall be calculated as the area within a perimeter projected trom the
perimeter of the wedge plate through the bearing plate at a 45 degree angle.
The behavior of the anchorage zone is controlled by the concrete strength and
by the reinforcement. The layout of the reinforcement and the tensile capacity
have a significant influence on the ultimate capacity and on its behavior at service
state. Different failure modes can occur, either in the local or in the general zone.
Failure in the local zone occurs in the immediate vicinity of the anchorage device.
The surface of rupture is often in the sharpe of a pyramid or cone, delimited by
.r
,..-,
'
245
_)
Burdet [42] indicates that single anchor configurations can be loaded concentrically or eccentrically as shown in Fig. 4.34. The geometry of the tendon can
be parallel to the axis of the anchorage zone or inclined, and also curved. Other
external forces like transverse post-tensioning or transverse external forces can
act on the anchorage zone.
__ j
'
a.concentric
b. eccentric
e. transverse post-tensioning
f. transverse reaction
Figure 4.34: Possible configuration for single anchor (from Ref. [42])
246
The different studies generally concentrated on the spalling forces and bursting
forces. The spalling forces are the tensile stresses acting in areas of the concrete
close to the end surface on either side of the anchorage device. These stresses
are essentially induced by the condition of compatibility of displacements. Guyon
recommended as a design va!ue 4% of the applied load as the corresponding
reinforcement in the form of a fine mesh, located as close to the face of the
concrete as possible. Burdet [42] shows this to be conservative. Since such
compatability induced forces cannot be determined from an equilibrium based
strut- and- tie model, the Guyon value is recommended for loaded-end face crack
control.
'
~,
'
247
The bursting stresses are the tensile stresses acting transversely to the axis of
the tendon at a certain distance ahead of the anchorage. Bursting stresses are
caused by the transverse spreading of the concentrated post-tensioning forces
overthe entire cross section. Figure 4.35 shows the geometry and nomenclature
for the simplest case, that of a concentric single anchor. Figure 4.36 shows some
comparison of Burdet's finite element Jnalysis [42] and Guyon's analysis. For
design purposes Guyon [125] and Leonhardt [129] presented the following
equation to compute the total bursting force:
T
y F (1 - a I d)
0.25 (Guyon)
0.30 (Leonhardt)
Many specifications have used similar equations. Good agreement with the
finite load analysis can be seen from Fig. 4.36.
0.3 .
.
0.2
Normalized
Bursting Force
Tburst I P
' . ..
.. ..
..
.
' ..
..'
'
....
..
+ FE-concentric
... . .
Guyon 0.3
...__ ........
4;. .. ..
.
0.1
- Guyon 0.25
.. ' ' .
.. . ..
.. ..
.....
... ' ...
-Guyon Analysis
...........,
~.
Figure 4.36: Comparison of finite element analysis with results from Guyon
(from Ref. [42])
248
Two different strut- and- tie- models are shown for the concentric single anchor
zone case in Fig. 4.37. Since the equations of equilibrium express the overall
equilibrium of the structure, both logically must give the same answer. The thrust
line model of Fig. 4.37(b) gives about 20 percent lowerstrain energy at ultimate
load level [42] than the simple strut- and- tie- model shown in Fig. 4.37(a). This
indipates more efficiency but is more related to the length of the transverse ties.
Since in actual detailing, the ties would be extended towards the outer
edges in both cases, this efficiency would not be practically developed and
either can be used.
249
... .,.
...
....
... ..................
.............
.........
"' '
..
..
++++-+.,
I
I
..............
+ +
'I I t
I I
I I
I
I
I
I
I
I t
I
I
I
I
I
t I
I
I
. .
I
''
I
I ' I I
I
I
I
I
I
I
' '
'
I
t
t
'
I
I
I
I
I
I
I
I
t t t
I ' I
I 4 ' I
t
I I
I I
I
I
I
I
..
t
I
I
I
I
I
I
Figure 4.37: Comparison of two different strut- and- tie- models with principal
stress vectors (from Ref. [42])
For design purposes the simple model of Fig. 4.37(a) can be used to
determine the required reinforcement as shown in Fig. 4.38. The
expressions for the angle of spreading of the compression force and the
location of the centroid of the reinforcement shown in Fig. 4.38 are based on
the results of Burdet [42]. The reinforcement should not be placed too close
to the anchor. The thrust line model of Fig. 4.37(b) forces the designer to
spread the reinforcement out more over the entire length of the D-region.
When using the simple strut- and- tie model the reinforcement should be
spread in a zone from 0.2 h" to 1.2h" [42]. Some additional transverse
reinforcement should be placed normal to these stirrups to resist the
spreading of the forces in the principal plane normal to this figure in the third
dimension (see example: '(
I # 3 bar).
250
f..__'d1 ~
~~reinforcement area=
1.0 h -
...,,..
,.__ _ _ _ _ _ Q.region : 1.5 h _ _ _ _ _ ___,.,....,...,
d1
= h 1 (4 tan a)
d1
Figure 4.38: Proposed strut- and- tie- model for anchorage zone
251
Example 4.5: Anchorage zone
Design the reinforcement required in the post-tensioning anchorage zone shown
in Fig. 4.39 to carry a maximum applied posHensioning force of 500 kips applied
to the centroidal axis of a beam with an overall height of 36 in. and a width of 12
in. The bearing plate is sufficiently rigid. It has a height of 8 in. and a width of 8
in. Assume concrete strength at time of stressing will be 6000 psi. Check both
local zone for proper confinement and general zone for both transverse bursting
reinforcement and spalling reinforcement. Both bars and spiral can be assumed
'2
as Grade 60.
=
=
=
=
=
=
500 kips
36 in.
12 in.
8 in.
8 in. (diameter of spiral)
1.5 in. (pitch of spiral)
Spalling Force:
<!J1
4>1
d1
=
=
=
=
12 + 3/(a/h) 05
18.4
(h- a) I (4 tan $1)
21.0 ln.
252
Internal forces:
=
=
=
C2
C2
=
=
=
Fl (2 cos <j>1)
263.5 kips
F 12 tan <j>1
83.2 kips
Fl2
250 kips
T1
T1
=
=
C1 sin <!>1
83.2 kips
C1
C1
co
co
As
83.2 kips
83.2 I 60
(Grade 60)
1.39 in 2
8- #4
1.60 in. 2
These bars must be distributed over a zone from 0.2 h (7.2 in.) to 1.2 h (43.2 in.)
.r
from the loaded face select #4 closed stirrups. One additional stirrup is located
as close to front face as cover requirements allow to provide required spalling
reinforcement. The 4 - #4 stirrups which satisfy the required 8 - #4 bars for T1
are then spaced at 8". This results in locations 10.5 in., 18.5 in., 26.5 in., and
34.5 in. as measured from the front face. One additional stirrup is provided at
42.5 in.
Out of plane:
A similar check must be made in the other principal plane. However, since the
=b = 12 in., there is appreFrom Fig. 4.36, with ajh 2 = 8112 =0.67, the
=
=
= 41.2 kips
= 0.69 in.2
=2.4 in.
=0.11
to 1.2 h2
in. 2 ). Six of
=14.4 in.
should
control such transverse splitting. These #3 ties are shown on Fig. 4.40. They
are also assisted by the spiral in this region.
!
253
;_j
36"
18"
_j
~-o~~lf-----d1
t-------D-region : 54"
1!-<ll
...
-,
: _j
Figure 4.39: Strut- and- tie- model for example: anchorage zone
..I
254
Anchorage requirements:
The #4 bars required for the T1 ties are adequately developed by bending them
around #4 longitudinal bars placed in the corners.
0.75
The critical node in anchorages is usually at the anchor plate. The geometry is
indicated in Fig. 3.8.
Check: CCC-node
Horizontal compressive stress in CCC-node
Since w4 > aO, hydrostatic stress
w4
a/2
w4
a/2
12 in.
()co
(jca
4 in.
= 4.5 ksi
OK
=
=
=
=
As
(318)2n I 4
fcon
=
=
(jca
fcon
d
fcon
[2 As fy I (d s)] (1 - sld) 2
8 in.
1.5 in.
726 psi
=
=
0.11 in. 2
0.726 ksi
(Assumes a #3 spiral)
L_,
255
(jca
7.79 ksi
OK
OK
Note that at this depth, y =a, the strut stress is essentially the same as the
stress at the end of the general zone
256
.,....._A
36"
L_j
f-ooll---12"
--l
#4 straight bars
-.3''
8"
36''
12"
l_
:......_-#4 stirrups
;.;.;;;;:.::::;:::::q
#3 ties
,...."'+--~;::~spiral
_L
8"
3"
2"1-- 8"'-12"1
-+--12"~
As a last example and in order to illustrate the application of strut- and- tie
models in both 'D' and '8' regions, a prestressed concrete beam wiH be analyzed
using the strut- and- tie model as shown in Section 2.5.1. Use of the strut- andtie model in B regions may be more cumbersome than ordinary sectional
analysis. By prestressing, forces are artificially created with the help of hydraulic
jacks. These forces act as loads on the prestressing stell and as loads on the
concrete. The designer chooses the tendon profile, the type and the magnitude
of the prestressing forces in such a manner that these artificial loads change the
internal force path created by the actual loads. Proper prestressing ordinarily can
prevent the formation of tension cracks under working loads. Also, the deflection
under working loads can be greatly reduced, because prestress usually puts
camber into a member under dead loading. A simple span beam is prestressed
by introducing a negative moment to offset the expected actual positive moment
and at the same time intoducing a longitudinal compression to offset the tensile
stresses from bending moment. Continuous beams are prestressed in a similar
fashion but for best results require an effective eccentricity above middepth in
negative moment zones. In ordinary computations, when continuous beams are
prestressed, secondary moments are introduced because the reactions prevent
full movement under the action of the prestress. One of the major advantages
of strut- and- tie models is that the prestress is introduced as forces acting on the
structure and their effect isdirectlyconsidered for the actual boundary conditions.
sscf
258
in the lower chord, this larger area corresponding to distributed strands will be used.
Further assume Md = 65 K-ft. and M1 = 80 K-ft.
For the uncracked section, the steel is transformed as (n-1) A 5 :
Atot
=
10*20+1.5(7-1)
=
209in. 2
y' from middepth of the concrete cross section
y'
-9 * 71209
10 * 20 I 12 + 200 * 0.32 + 9 * 6. 7 2
7087 in. 4
After cutting the tendons the compression and tension chord forces the strut-and-tie
model shown in Fig. 4.42(b) can be computed:
Pi
(1.5) (135)
ICP,
Pi I (2 cos 12)
202 kips
=
103.3 kips
In order to check the compressive stresses resulting from the application of this
concentrated force by the distributed strands, it is assumed that the centroid of the
strands is 3 in. from the bottom and that they are fully distributed over the width.
ThusAP1 = (2) (3) (10) = 60in. 2
For f
c'
For f j =
ve
= 5000,
fj = 0.75 * 4000
ICP 1 I AP 1 =
103.3 I
= 3000 psi
60 = 1. 72 ksi
< 0. 75 * 4000
= 3.0
ksi
OK
L.
Of substantial concern is the need for lateral and vertical reinforcement throughoutthe
transfer length to resist the tension forces Tp1 shown in Figure 4.42(a).
ITp1 =
As
=
Note that these tension forces exist laterally as well as vertically so that only 1 leg on
the bottom of each stirrup runs transversely to resist the lateral component.
As for each stirrup is 0.11 si.
Thus
,----
259
If the 'D' region at the end is isolated as shown in Fig. 4.42(c) and the combined
stresses due to the prestress and its eccentricity are computed from
PIA+ Pee /I , the values given in Fig. 4.42(c) are found. Applying these
stresses as forces T1 = C2 and C3 =Pi at their respective centroids as indicated, it is very easy to construct the force path and strut- and- tie model shown.
This clearly illustrates 'that if tensile strength of concrete is not to be relied on, an
area of steel As= 26/60 = 0.43 si should be provided in the end regions close
to the top of the beam. Two #4 bars are provided as shown in Fig. 4.43. They
also are useful for positioning and. anchoring the stirrups. This 'D' region also
indicates the need for a similar area of vertical reinforcement at the support. The
closely spaced #3 stirrups provided over the support to work locally to resist
strand splitting forces also work nicely over the full depth to provide this resistance. The advantage of strut- and- tie modelling in the 'D' regions is clear from
these types of calculations.
--,
'
260
I= 51ft
wdl+tl = 0.446 k I ft
M
dl
II
Section: A - A
17"
20"
_j
3"
'
~10"~
~10"~
Atot = Ac + (n-1) Ap
2
261
Tp1 a (P tan ~1 ) 12
,- -,
.,.Cp1
.._.....
...
I
I
I
I
.,., ,---cp2-;-
ftllll'
.,.
ftllll'
: :.,. .,.
ftllll'
~ .....
...
------.,. .,.
.,.
.,.
.,. .,.
.._ .._
.....
. . . .....
----~~-----~~-
.I
(a)
r
h
L
Tp = wt b (-PI Ar_ + P e Yt/IJ
(b)
262
T1
C2
I
I
I
ft-1.00
\.j
\,
\
Q>T
\,
',
C2= 26k
',
,1, __
,---------------------------~I<~~
T = 26k
IT~
1<----\
.
I
I
I
I
I
C3 =202
I
k
I
I
I
I
I
k----\
k-----\
~~ ~-------\
-Ei A
!l~ ~~-----,--~----'
-\
-2.82
!
\
centroid of compression
force outside shaded triangle
Choice is arbitrary
as long as angles
do not get to small.
In this case, chosen so
<I>T = 45
(c)
263
Top chord: tension from prestressing forces after cutting the wires
'-~
It can be seen from Fig. 4.42(a) that the prestress force is applied to the lower
chord gradually over the transfer length, 50 db= 25 in. It is assumed that this
can be approximated as three equal forces, Pi, each located about 8 in. apart.
However, when compared to the length of the beam, the critical zone for tension
on the top chord can be effectively checked with full prestress and no dead load
moment. This is slightly severe but practical. The prestressing loads are applied
to the overall beam as any other load would be. This load case illustrates one
problem with strut- and- tie modelling. As previously shown in Fig. 2.30(d), if a
simple truss model is used assuming free articulation at all joints, the application
of a horizontal force concentric with the centroid of the lower chord to a simply
supported, articulated, simple truss does not produce top chord forces. However,
in anormal mechanics analysis, it is assumed that plane sections remain plane
so that the conditions of deformation compatibility are introduced. These are not
part of an equilibrium or plastic analysis. Thus in the highly elastic prestressed
beam at service load conditions, these compatibility considerations are necessary and some beam analysis concepts must be introduced.
=
The effective top chord depth, W1 , is estimated as having a centroid about as far
from the outer fiber as the centroid of the strands, 3 in. Since the beam has
uniform width and since the final stress distribution is assumed uniform, W1 =6 in.
and the distance from the section centroid to the centroid of this chord is 10.33 = 7.3 in. (One can see that lumping of all top chord fibers into a single chord
reduces the accuracy of outer fiber stress calculations done in the traditional
PIA+ Mc/1 manner.)
=
rd
60.0 kips
264
Bottom chord:
169.1 kips
Check end section where dead load moment is zero under effect of prestress forces:
Tog chord:
Tp = 60 kips~ 6v
N.G.
Cp = 169.1 kips~ 0. 75
z
cdl
=
=
=
Tdt
=
Md, I z
14.3 (from Ref. [55], or as a first approximation: z = 314 h)
Td1
=
780/14.3
=
54.55 kips
Check centerline section under effect of dead load combined with prestress forces:
Top chord:
T (prestr .} - C(load}
60- 54.55- 0
<
=
=
3 v'
f:
At (tension}
11.38 kips
Bottom chord:
T (load) - C(prestr.) ~- b wt
OK
/6 tt'
use of
OK
f: as well.
265
=
Mdl+n/ z
14.3 (from Ref. [55], or as a first approximation: z = 3/4 h)
cd,+ll
Td,+n
Td,+ll
1740/14.3
121.7kips
266
top chord:
I
T {prestr.)- C{load) ~- b W 0 v fc
60 * 149.5 I 202 - 121.7 = -77.6 {compression)
~-
10 * 6 * 0. 75 * 5
tension chord:
C(prestr.) - T(load) ~
6 in.
b wt 3 (f'c )0.5
2 in.
=3.45 (tension)
i 0 * 2 * 3 * (5000) 0 5 =4.24 kips
(tension)
OK
1111111111 Fllllllllllll
1 ------------,
-- - -1
I
/ I
vertical chord:
Tv
Tv
Tv
=
=
=
"4so
Ll
I
A-w*x
//
~------d:J
<
T
#3 U stirrup = 2 * 0.11 * 60 = 13.2 kips
267
Obviously the numbers and reinforcement would change if load factors and
resistance factors were applied. However, the principles would be the
same. It can be seen that the allowable stress checks using strut-and-tie
models are more cumbersome than conventional sectional analysis
procedures.
In order to illustrate the check for ultimate conditions, assume ACI
Building Code load factors and q> = 1.0. Then Mu = 1.4(780) + (1.7)(860) =
2724 in.k.
Cu
= T u. =
Mu I z
= 2724 I
14.3
condition is truly plastic so the basic strut- and- tie model applies. Checking
at the centerline with an effective prestress of 149.5 kips:
C (load)
Tog Chord
190.5
<
<
b We Ve fc
( 10) (6) (. 75) (5)
= 225
Bottom Chord
T (load)
190.5
<
<
T + p
(1.5) (250)
OK
=
(APfPv - Pe)
375 kips
APfPY - P = AP (fPY- fP) is the reserve capacity in the tendons after prestressing.
This is the tension that can be developed in the tendons above that developed during
prestressing.
p
OK
268
1:
Again, in reality, proper load factors and <j> factors would be required. For definitions refer to Fig. 3.12. In this case there would be 3 layers of 112 .. strands, each
2 in. on centers. Thus wT = 3(0.5) + (2)(1.5) = 4.5 in. The inclined compression struts were assumed at an angle <j> = 45. C1 =A= (51 )(0.446)(112} = 11.37
kips. Thus C2
8 in.
wT
4.5 in.
w2
(j'c2
C21(w2*b)~Vefc
(j'c2
()a
()a
11.37 I (8 * 10)
OK
OK
:
[
__
269
12~--
r
l
20"
t-A
4@5"
.. )
I
.
'
..
''
::::::
1::::;::
i::':
'
I?
:=::;;::
:::
t:
1.1.1. pol.l.ll
~10"4
::::
7:
t':::;:::
::-:
~15"~
:;:;::
::::
~;:;;:;:;
:;:::
:::::::;:;:;::
:::::
;::::::
.:::::::::::::
~r
I
I
... A
Section: A- A
#4
2"
iollt----10"
-----1
'
:;::
i~
:~l~'
'}::::::
''
270
_)
o or discontinuity regions.
While it can be
used in the more regular s or Bernoulli regions where linear strain profiles are
encountered, it is not as advantageous as ordinary structural concrete design
procedure in those regions. After discussing the general principles, components
and modeling techniques as well as dimensioning of the struts, ties and nodes
in chapters 2 and 3, illustrative design examples were presented in chapter 4.
Several typical strut- and- tie patterns are furnished in Appendix A.
For the majority of concrete structures it would be unreasonable and
inefficient to model the entire structure with a strut- and- tie- model. It is
advantageous to subdivide the given structure into B-regions and D-regions.
After computing the elastic stress resultants or ordinary cracked reinforced
concrete forces for the B-regions, the equivalent forces should be applied to the
D-regions. Load paths can be sketched based on experience, design aids,
experimental results or a finite element analysis. The strut- and- ties can be
rearanged with consideration of practicality of the reinforcement layout. The
proposed design recommendations are applicable to either prestressed or nonc
271
272
In this report, the design procedures based on the strut- and- tie- model
and the proposed detailing approaches are illustrated with a series of design examples. In addition, several strut- and- tie- models (from Ref. [2, 28]) which may
be useful to the designer when detailing 'D' regions in concrete structures are
included in Appendix A.
Study of the design examples indicates that use ofthe strut- and- tie model
is an extremely efficient way of detailing reinforcement in 'D' regions. The
calculations are relatively simple and straightforward and give the designer
substantial insight. In contrast, the checks of struts and nodes are laborious and
somewhat subjective. It was noted that in many applications these strut and node
stresses were not close to controlling design. Hopefully, further application and
familiarity with the method will give designers a "feel" for when detailed strut and
node calculations are required and when they can be assumed as not governing.
L ..
6.
References
1.
Schlaich, J.; Weischede, D.: A practical Method for the Design and
Detailing of Structural Concrete (in german). Bulletin d'lnformation no.
150. Comite Euro- International du Beton, Paris, 1982 163 pp.
2.
3.
4.
5.
6.
7.
Barton, D.: Design of Dapped Beams using the Strut- and- Tie- Model.
Master's Thesis, The University of Texas at Austin, 1988
8.
Withey, M.:Tests of Plain and Reinforced Concrete, Series of 1906 and 1907.
Bulletin of the University of Wisconsin, Engineering Series, Vol. 4, No.2,
1907- 1908
9.
10.
11.
\_
-'
273
274
12.
13.
14.
15.
16.
17.
18.
Grab. J.: Ultimate Strength of Beams with thin walled open crosssections. Bericht 56, lnstitut fOr Baustatik und Konstruktion-ETH, ZOrich,
1970
19.
20.
I'
--,
'
'"'
;-%
275
'
21.
Nielsen, M.; Braestrup, N.; Jensen, B.; Bach, F.: Concrete Plasticity:
Beam Shear-Shear in Joints -Punching Shear. Specialpublikation Dansk Selskab for Bygningsstatik, Lyngby, 1978, 129 pp.
22,
23.
Ramirez, J.; Breen, J.: Proposed Design Procedures for Shear and
Torsion in Reinforced and Prestressed Concrete. Research Report 2484F, CenterforTransportation Research-The University of Texas at Austin,
1983, 254 pp.
24.
25.
Dei, Poli, S.; Gambarova, P.; Karakoc, C.: Discussion on the papers
"Shear Transfer across cracks in R.C. due to Aggregate Interlock and to
Dowel Action (MRC, No. 126} and "Shear Transfer in R. C." (MCR, No.
130) by Millard, S. and Johnson R., Magazine of Concrete Research, Vol.
38, No. 134, 1986, pp. 47-51
26.
27.
28.
29.
30.
276
31.
32.
33.
Hartmann, D.; Breen, J.; Kreger, M.: Shear Capacity of High Strength
Prestressed Concrete Girders. Research Report 381-2. Center for
Transportation Research-The University of Texas at Austin, 1988, 244 pp.
34.
35.
36.
37.
38.
Anderson, R.: Behavior of CTT-Nodes in Reinforced Concrete Strut- andTie- Models. Master's Thesis, The University of Texas at Austin, 1988
39.
Bouadi, A.: Behavior of CCT-Nodes in Structural Concrete for the Strutand- Tie- Model. Master's Thesis, The University of Texas at Austin,
1989
40.
Noguchi, H.; Watanabe, K.: Shear Resistance Mechanisms of BeamColumn Joints under Reversed Cyclic Loading. In: IABSE Colloquium,
Delft, 1987, pp. 511-522
41.
42.
(_
\_
277
43.
44.
45.
46.
47.
48.
Soroushian, P.; Obaseki, K.; Rojas, M.; Sim, J.: Analysis of Dowel Bars
acting against Reinforced Concrete Core. ACI -Journal No. 83-59,
1986, pp. 642-649
49.
Vintzeleou, E.; Tassios, T.: Mathematical Models for Dowel Action under
Monotonic and Cyclic Conditions. Magazine of Concrete Research, Vol.
38, No. 134, 1986, pp. 13-2~ 45.
50.
51.
52.
53.
54.
Richart, F.; Brandtzaeg, A.; Brown, R.:A Study of the Failure of Concrete
under combined Compressive Stresses. Bulletin No. 185, University of
Illinois, Urbana, 1928, pp. 7-103
278
55.
56.
ThOrlimann, B.; Marti, P.; Pralong, J.; Ritz, P.; Zimmerli, B.: Anwendung
der Plastizitatstheorie auf Stahlbeton (Plasticity in Concrete Structures}.
Vorlesung, ETH- ZOrich, 1983
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
L_,
279
67.
68.
Kupfer, H.; Hilsdorf, H.; ROsch, H.: Behavior of Concrete under Biaxial
Stresses. ACI-Journal, Vol. 66, No.8, August 1969, pp. 656-666
69.
70.
71.
72.
Han, D.; Chen, W.: A Nonuniform Hardening Plasticity Model for Concrete
Materials. Mechanics of Materials, No. 4, 1985
73.
74.
Fardis, M.; Alibe, B.; Tassoulas, J.: Monotonic and Cyclic Constitutive Law
for Concrete. Journal of Engineering Mechanics, Vol. 109, No. 2, 1983,
pp. 517-536
75.
76.
77.
78.
79.
... J
omig, K.: A Proposal for a Draft Code of Practice for Prestressed Concrete.
Cement and Concrete Association, London 1948
81
82.
83.
84.
85.
86.
Nijogi, S.: Concrete Bearing Strength- Support, Mix, Size Effect. Journal
of the Structural Division, ASCE. ST8, 1974, pp. 1685 - 1702
87.
88.
89.
Roberts, C.: Behavior and Design of the Local Anchorage Zone in PostTensioned Concrete. Master- Thesis, The University of Texas at Austin,
1990
90.
91.
Ramirez, J.; Breen, J.: Review of Design Procedures for Shear and
Torsion in Reinforced and Prestressed Concrete. Research Report 2482, Center for Transportation Research-The University of Texas at Austin,
1983, 186 pp.
'--'
281
92.
93.
94.
95.
Hoyer, E.; Friedrich, E.: Beitrag zur Frage der Hafspannung in Eisenbeton bauteilen, (Contribution to Questions Regarding Bond Stresses in
Reinforced Concrete Structural Members). Beton und Eisen, Vol. 38, No.
6, 1939, pp. 102-110.
96.
97.
98.
99.
Jirsa, J.; Lutz, L.; Gergely, P.: Rationale for Suggested Development,
Splice, and Standard Hook Provisions for Deformed Bars in Tension. In:
Concrete International, Vol. 1, No.7, 1979, p. 47
100.
Orangun, C.; Jirsa, J.; Breen, J.: Reevalauation of Test Data on Develop
ment Length and Splices. ACI-Journal, 74, No.3, 1977, p. 114
101.
102.
103.
282
104.
105.
Park, R.; Paulay, T.: Reinforced Concrete Structures. Wiley-Interscience, New York, 1975, p. 671
106.
107.
108.
109.
110.
111.
112.
113.
114.
115.
L___:
283
116.
117.
118.
119.
Peterson, R.: Stress Concentration Factors. New York. 1974. pp. 208
120.
121
122.
Shah, S.N.: Evaluation of infilf wall strengthening schemes for nunductile reinforced concrete buildings. Master's Thesis, The University
of Texas at Austin, 1989
122.
124.
Leonhardt, F.: Vorlesungen Ober Massivbau. Teil1 bis Teil6. SpringerVerlag, Berlin, 1984
125.
126.
Magnel, G.: Prestressed Concrete. New York, Third edition 1954, 345 pp.
127.
Lenschow, R.; Sozen, M.: Practical Analysis of the Anchorage Zone Problem
in Prestressed Beams. In: ACI- Journal, November 1965, pp. 1421 - 1437
128.
' 'J
' .. )
' J
'~_;
284
,.,
129.
130.
131.
Barton, D.L.; Anderson, A.B.; Bouadi, A.; Jirsa, J. 0.; and Breen, J.E.: "An
Investigation of Strut-and-Tie Models for Dapped Beam Details," Research
Report 1127-1, Center for Transportation Research, The University of Texas
at Austin, May 1991.
I_
L;
~-,
Appendix A
Detailing Aids
(from Ref. [1) and [2))
_....._
___________
.-
285
286
C1
= F I (2 cos $1)
T2 = CO = F /2 tan $1
$1
=12 + 3/ "(all)
287
C1
= F I cos q>1
T1
=C2 = F tan ~1
~,
=12 + 3/"(a/1)
'
I.
--Tz
_l
~~,
d
I
~~
I~
.I
1- \
C1 = F I (2 cos cp1)
CO = F I 2
tan lj>1
T2 + C2 = F
tan 4>1
lj>1 = 12 + 31 ..J(a/1)
Fo r d/1 :::;; 1: T1 = T2 /3
Fo r d/1 ~ 2: T1 = T2 /2
\. __
. _i
C1
=F I (2 cos cp1)
CO =F /2 tan cp1
T2 + C2
4>1
=F tan 4>1
= 12 + 3/ "'(all)
For d/1 s; 1: T1
=T2 I 3
~ 2: T1
=T2/2
For d/1
290
2F
d ~I
L_
~a~
1-CO= F tan
T1
=F tan e
e)
1 = 12 + 3/.f(all)
r
I
l.
L.
291
co
C1 ~
/
.~
T2
I
I
''
'
~-C1
I
I
I
I
C2
I
d >I
/L--c3---~
~
/
C4
C4
'r;,-
i.f
CO = C3 = F tan 4> 1
T1
C1
C2=F
. J
4>1
=12 + 3/l{all)
T1
''
'
t
F
:n
T1
0.313F
1.375F
~a~
~a~
.I
sI
~~.
0.313F
T1 = 1.375 F tan
T2
co= 0.313 F
0.313 F
Ji
1.375 F
C2 = 0.16 F I (cos 6 )
C3 = 0.69 F I (cos $1 cos 6)
T3 =1.062 F tan a
.JL
C1
tan
C4
=0.69 F I (cos 6)
4>1
=12 + 3/.f(an)
0.313 F
tlllllllllllllllllllllllllllllll l!!llllllllllllllllllllllllllllllllt
I
C3
C3
J---
I
C2- -
I
C1
C1
a,
VB
\
~.
:n
a =1/ (4 z) "" 1/ (3 d)
F=C3= wl/2
=F
z = 3/4 d
r--1
T1
-~
\
arctan
tan
293
C1 = F /cos
C2=T1 =Ftane
r-1/4
..
294
1.
r----C2---
I
I
~
I
\
C1
C1
T2
T2
z = 314 d
\
'9"
\
\
I
T1
te lllllllllllllllllllllllllli11111111111111111111111111 ~
F
laJ
tu4~
arctan 6 =II (4 z) =II {3d)
F =C3 = w 112
T1 = F tan 6
C1
T2=F
F I cos B
.I
----''
---
.. J
--T2
T1
_j
''
_c1
' '
T2
T2
''
'
,.
'C2
'
'
'
!!;)-.
-- ...
~th
......
t
L
r-b
hO
_l
..
1---bo--
Tt = Mt/ (2 bO ho)
=hO /6
tb = bO /6
Ts= S /(2 z)
T1
=T /tan b
T1 =T
=T
for 45
j
for 45
T2
I
~th
-. ~
r
th
st
T2=T
T I (hO sin
tan b
ocos b) :S v f~
295
''
'
296
L. ___ _
=C1 *z
T3 = T1 * tan a2
C1 =T1
T3 = 0.3 T1
C1 I 2 * w1 * b ~ v f(;
T4
=T1 *
=T1 I cos a2
a2
=15-18 o
C2
297
. 1
_j
I __ j
. i
_j
---,
.J
_j
. J
M = C1 * z
C1 =T1
C1/ 2 * w1 * b
C2 = C1/sln
l_j
T2 = T1/sln 8
_j
v f~
C2/2*w2"b.:S; vf~
C3 = C1/tan
:S;
298
co
/
C1
/T1f T1~' ,
F
/
~A~
.~T2
'
j
'
;;,
T2~
C31 C3
~ ~1
I
I
C2
C1
I
T3
~
----,
\.
.
I
C2
I
I
I
I
I
C4
I
I
C4
..
I
I
tlllllllllllllllllllllllll*: lllllllllllllllllllllllllt
I
T1
.J
=F 14
CO= F/4
T2 = F 14
C1 =2 F 1/2
T3 = F I 4 tan 1
C2=F I4
C3 =F I 4 cos 4>1
C4=F I4
w6~
299
~aF
' l
1.5 d
."
..,1
I_
z = 0.75 h
arctan
X: 0.75
w6 ;;:: C6 I (2 b v ~ )
91:12+3/-J(a/f)
T2
=F I tan (B+91)
C1
= F I sin (B+91)
T1
T5:2F
C6:F
300
r ,
I
l
=2 C1 .. 2 cos 45 T4
C1 :T1
C6
T2 = T1 tan y
for r= 15
T2 T1 * 0.268
T7
=
T3 =C1 * tany I
for r= 15
T3 =C1 * 0.379
cos 45
T5 = C1 I cos 45. T4
for r= 15
T5 = C1 * 0.896
=2 C6 .. 2 C11
T8 =0.707 T9
C1 I (b * 2 * w1) s; v ~
C11 I (b * 2 * w11) s; v ~
,.